Safety Assessment of Trimellitic Anhydride Copolymers As Used in Cosmetics

Safety Assessment of Trimellitic Anhydride Copolymers as Used in Cosmetics

Status: Draft Report for Panel Review Release Date: August 28, 2015 Panel Meeting Date: September 21-22, 2015

The 2015 Cosmetic Ingredient Review Expert Panel members are: Chairman, Wilma F. Bergfeld, M.D., F.A.C.P.; Donald V. Belsito, M.D.; Ronald A. Hill, Ph.D.; Curtis D. Klaassen, Ph.D.; Daniel C. Liebler, Ph.D.; James G. Marks, Jr., M.D., Ronald C. Shank, Ph.D.; Thomas J. Slaga, Ph.D.; and Paul W. Snyder, D.V.M., Ph.D. The CIR Director is Lillian J. Gill, D.P.A. This safety assessment was prepared by Monice M. Fiume, Assistant Director/Senior Scientific Analyst/Writer and Bart Heldreth, Ph.D., Chemist.

Commitment & Credibility since 1976

Memorandum

To: CIR Expert Panel Members and Liaisons From: Monice M. Fiume MMF Assistant Director/Senior Scientific Analyst Date: August 28, 2015 Subject: Safety Assessment of Trimellitic Anhydride Copolymers as Used in Cosmetics

Enclosed is the Draft Report on the Safety Assessment of Trimellitic Anhydride Copolymers as Used in Cosmetics. (It is identified as melply092015rep in the pdf document.) This is the first time the Panel is seeing a safety assessment on this family of six ingredients. The Scientific Literature Review was issued on June 1, 2015.

The trimellitic anhydride copolymers included in this report form a family of cosmetic ingredients because that they all share in common trimellitic anhydride (aka 1,2,4-benzenetricarboxylic acid anhydride) as a monomer. Each copolymer is also composed of 1 to 4 of the following additional monomers: adipic acid, cyclohexanedimethanol (CHDM), ethylene glycol, isophthalic acid, maleic anhydride, neopentyl glycol, phthalic anhydride, propylene glycol, sebacic acid, trimethylpentanediol, or trimethylolpropane (chain-terminated by isostearic acid). Most of these ingredients are reported to function in cosmetics as film formers.

No information has been submitted regarding the amount of each residual monomer present in these copolymer ingredients, and therefore it should be assumed that it is possible that residual amounts of the monomer(s) could be present in each copolymer. Accordingly, relevant information on the toxicity of these monomers is presented in Table 2 of the report. Much of this information came from ECHA or OECD SIDS reports, and some from previous CIR reports (melply092015prev_1 through melply092015prev_5.) This information is not intended to be exhaustive or complete, but purely summary information intended to provide insight as to any possible toxicity concerns for these monomers.

Concentration of use data (melply092015data) were submitted by the Council. Comments on the SLR (melply092015pcpc) also were received from the Council and have been addressed.

If the data included in this report adequately addresses the safety of the trimellitic anhydride copolymers, the Panel should be prepared to formulate a tentative conclusion, provide the rationale to be described in the Discussion, and issue a Tentative Report for public comment. If the data are not sufficient for making a determination of safety, then an Insufficient Data Announcement should be issued that provides a listing of the additional data that are needed.

SAFETY ASSESSMENT FLOW CHART

INGREDIENT/FAMILY _____Trimellitic Anhydride Copolymers______

MEETING ______Sept 2015______

Public Comment CIR Expert Panel Report Status

Priority List INGREDIENT PRIORITY LIST

SLR

June 1, 2015

DRAFT REPORT Draft Report Sept 2015 60 day public comment period

Table

Table IDA TR

IDA Notice IDA

DRAFT TENTATIVE 60 day public comment period Draft TR REPORT

Table

Tentative Report Issue TR

Draft FR DRAFT FINAL REPORT

60 day Public comment period

Table

Table Different Conclusion

PUBLISH Final Report Issue FR

Distributed for comment only -- do not cite or quote

Trimellitic Anhydride Copolymers – Sept 2015 – Monice Fiume

icity

Acute, Acute, Rptd Rptd

– – – – al ation

ation

Use Reported Mfg of Method Impurities Toxicokinetics Dermal Penetration Animal Tox Dermal Animal Tox Oral Animal Tox, Acute, Inhal Animal Tox Derm Dose, Animal Tox, Rptd Dose, Oral Animal Tox Dose, Inhal Tox Repro/Dev Genotox icity Ocular Irritation Irr Membrane Mucous Dermal Irr/Sens Reports Case Carcinogenicity

General Trimellitic Anhydride X X Copolymers Adipic Acid/CHDM/MA/ Neopentyl Glycol/Trimellitic Anhydride Copolymer Adipic Acid/Neopentyl Glycol/ X Trimellitic Anhydride Copolymer Isostearoyl Trimellitic Anhydride/ Trimethylolpropane Copolymer Phthalic Anhydride/Trimellitic X X X Anhydride/Glycols Copolymer Propylene Glycol/Sebacic Acid/

Trimellitic Anhydride Copolymer Trimethylpentanediol/Isophthalic

Acid/Trimellitic Anhydride Copolymer

Distributed for comment only -- do not cite or quote

INFORMATION ON MONOMERS

Dose,

icity

Acute, Rptd Acute, Rptd Rptd

. .

CIR by Reviewed Toxicokinetics Dermal Absorption Animal Tox – Dermal Animal Tox, Rptd Oral Animal Tox – Dose, Inhal Tox Repro/Dev Genotox icity Ocular Irritation Irr Membrane Mucous Animal Tox – Oral Animal Tox, Acute, Inhal Respiratory Effects - Human Recommended Limits Exposure Carcinogenicity Dermal Irr/Sens Animal Tox – Derm Dose, trimellitic anhydride X X X X X X X X X X X X adipic acid X X X X X X X X X X cyclohexanedimethanol X X X X X X X ethylene glycol X X isophthalic acid X X X X X X X X X X maleic anhydride X X X X X X X X X X X X X neopentyl glycol X X X X X X X X X phthalic anhydride X X X X X X X X X X X propylene glycol X X X X X X X X X X sebacic acid X X X X trimethylpentanediol X X X X X X X X trimethylolpropane X X X X X X X X

Distributed for comment only -- do not cite or quote

Trimellitic Anhydride Copolymers

Ingredient VCRP InfoBase PubMed SciFinder ChemPortal NTIS FDA ECHA HPV IUCLID/SIDS EU NICNAS Web Adipic Acid/CHDM/MA/Neopentyl Glycol/ 0 --- 0 0 0 0 0 0 no 0 Google Trimellitic Anhydride Copolymer restriction ChemID

No hits: PubChem HSDB ChemSpdr Adipic Acid/Neopentyl Glycol/Trimellitic 411 CAS#: there/nothing 0 0 0 0 0 no not assessed Google; Anhydride Copolymer supplier useful restriction PubChem; (CAS # 28407-73-0) ChemID Technical Names: Adipic Acid, Polyester with 1,2,4-Benzenetricarboxylic Acid Cyclic 1,2- No hits Anhydride and 2,2-Dimethyl-1,3-Propanediol; HSDB by CAS # - Hexanedioic Acid, Polymer with 1,3-Dihydro- ChemSpdr refined by 1,3-Dioxo-5-Isobenzofurancarboxylic Acid and document type – 2,2-Dimethyl-1,3-Propanediol; 6 hits/0 useful 1,2,4-Benzenetricarboxylic acid, cyclic 1,2- anhydride, polyester with adipic acid and 2,2- trimellitic dimethyl-1,3-propanediol; anhydride 1,3-Propanediol, 2,2-dimethyl-, polyester with copolymer – 286 adipic acid and 1,2,4-benzenetricarboxylic acid hits/4 useful cyclic 1,2-anhydride ; 5 hits/5 used 1,3-Propanediol, 2,2-dimethyl-, polymer with 1,2,4-Benzenetri- 1,3-dihydro-1,3-dioxo-5-isobenzofurancarboxylic carboxylic acid and hexanedioic acid; anhydride – 68 5-Isobenzofurancarboxylic acid, 1,3-dihydro- hits/4 useful 1,3-dioxo-, polymer with 2,2-dimethyl-1,3- propanediol and hexanedioic acid; nothing new Adipic acid-1,2,4-benzenetricarboxylic searching by anhydride-neopentyl glycol copolymer; individual names Adipic acid-1,2,4-benzenetricarboxylic anhydride-neopentyl glycol polymer; Adipic acid-neopentyl glycol-trimellitic anhydride copolymer; Adipic acid-neopentyl glycol-trimellitic anhydride polymer Supplier/Trade Names: Uniplex 670-P (Unitex Chemical Corporation) Isostearoyl Trimellitic Anhydride/ 0 CAS#: 0 0 0 0 0 0 no 0 No hits: Trimethylolpropane Copolymer supplier restriction PubChem (CAS # 1190965-82-2) ChemID Supplier/Trade Names: Schercemol ISTC Ester HSDB (Lubrizol Advanced Materials, Inc.) ChemSpdr; Google Distributed for comment only -- do not cite or quote

Ingredient VCRP InfoBase PubMed SciFinder ChemPortal NTIS FDA ECHA HPV IUCLID/SIDS EU NICNAS Web Phthalic Anhydride/ Trimellitic Anhydride/ 74 CAS#: 0 0 0 0 0 0 no 0 No hits: Glycols Copolymer supplier; restriction PubChem (CAS # 186688-25-5) TRN 20 (7) ChemIDHS Technical Names: Phthalic 2006 DB Anhydride/Trimellitic ChemSpdr; Anhydride/Glycol/Neopentyl Glycol Google Copolymer; Phthalic/Trimelitic/Glycols Copolymer; 5-Isobenzofurancarboxylic acid, 1,3-dihydro-1, 3-dioxo-, polymer with 2,2-dimethyl-1,3-propanediol, 1,2-ethanediol and 1,3-isobenzofurandione; 1,2-Ethanediol, polymer with 1,3-dihydro-1,3- dioxo-5-isobenzofurancarboxylic acid, 2,2- dimethyl-1,3-propanediol and 1,3-isobenzofurandione; 1,3-Isobenzofurandione, polymer with 1,3- dihydro-1,3-dioxo-5-isobenzofurancarboxylic acid, 2,2-dimethyl-1,3-propanediol and 1,2- ethanediol; 1,3-Propanediol, 2,2-dimethyl-, polymer with 1,3-dihydro-1,3-dioxo-5-isobenzofurancarboxylic acid, 1,2-ethanediol and 1,3- isobenzofurandione; Ethylene glycol-neopentyl glycol-phthalic anhydride-trimellitic anhydride copolymer Supplier/Trade Names: Polynex Resin (Estron Chemical, Inc.) Propylene Glycol/Sebacic Acid/Trimellitic 0 --- 0 0 0 0 0 0 no 0 No hits: Anhydride Copolymer restriction PubChem ChemID HSDB ChemSpdr Trimethylpentanediol/Isophthalic 0 --- 0 0 0 0 0 0 no 0 No hits: Acid/Trimellitic Anhydride Copolymer restriction PubChem ChemID HSDB ChemSpdr

Distributed for comment only -- do not cite or quote

SciFinder – Feb 24 2015: trimellitic anhydride copolymer (refined by document type) – 286 hits/4 useful Search by CAS # (28407-73-0; 1190965-82-2; 186688-25-5) (refined by document type) – 6hits/0 useful 1,2,4-Benzenetricarboxylic anhydride – 68 hits/4 useful Searched by individual names – nothing additional found Set up Keep Me Posted alerts for all SciFinder search strategies

PubMed – Feb 24 2015: (((((((((Adipic Acid/CHDM/MA/Neopentyl Glycol/Trimellitic Anhydride Copolymer) OR Adipic Acid/Neopentyl Glycol/Trimellitic Anhydride Copolymer) OR 28407-73-0[EC/RN Number]) OR Isostearoyl Trimellitic Anhydride/Trimethylolpropane Copolymer) OR 1190965-82-2[EC/RN Number]) OR Phthalic Anhydride/Trimellitic Anhydride/Glycols Copolymer) OR 186688-25-5[EC/RN Number]) OR Propylene Glycol/Sebacic Acid/Trimellitic Anhydride Copolymer) OR TDI/Trimellitic Anhydride Copolymer) OR Trimethylpentanediol/Isophthalic Acid/Trimellitic Anhydride Copolymer – 4 hits ((28407-73-0[EC/RN Number]) OR 1190965-82-2[EC/RN Number]) OR 186688-25-5[EC/RN Number] – 0 hits Trimellitic AND Anhydride AND Copolymer – 5 hits/5 useful 1,2,4-Benzenetricarboxylic anhydride – 7 hits/0 useful ((trimellitic OR benzenetricarboxylic) AND anhydride) AND (irrita* OR sensiti*) – 202 hits/17 useful

Distributed for comment only -- do not cite or quote

Safety Assessment of Trimellitic Anhydride Copolymers as Used in Cosmetics

Status: Draft Report for Panel Review Release Date: August 28, 2015 Panel Meeting Date: September 21-22, 2015

Distributed for comment only -- do not cite or quote

INTRODUCTION This assessment reviews the safety of the following 6 trimellitic anhydride copolymers as used in cosmetic formulations: • Adipic Acid/CHDM/MA/Neopentyl Glycol/Trimellitic Anhydride Copolymer • Adipic Acid/Neopentyl Glycol/Trimellitic Anhydride Copolymer • Isostearoyl Trimellitic Anhydride/Trimethylolpropane Copolymer • Phthalic Anhydride/Trimellitic Anhydride/Glycols Copolymer • Propylene Glycol/Sebacic Acid/Trimellitic Anhydride Copolymer • Trimethylpentanediol/Isophthalic Acid/Trimellitic Anhydride Copolymer

Most of the trimellitic anhydride copolymers are reported to function as film formers in cosmetic formulations1 (Table 1). The ingredients in this report are related as copolymers in that they all share in common trimellitic anhydride (aka 1,2,4-benzenetricarboxylic acid anhydride) as a monomer. Each copolymer is also composed of 1 to 4 of the following additional monomers: adipic acid, cyclohexanedimethanol (CHDM), ethylene glycol, isophthalic acid, maleic anhydride, neopentyl glycol, phthalic anhydride, propylene glycol, sebacic acid, trimethylpentanediol, or trimethylolpropane (chain-terminated by isostearic acid).1 However, no information has been submitted regarding the amount of residual monomer present in each copolymer, and therefore it should be assumed that it is possible that the residual monomer(s) can be present in the copolymer. Accordingly, relevant information on the toxicity of these monomers is presented in Table 2.2-31 This information is not intended to be exhaustive or complete, but purely summary information intended to provide insight as to any possible toxicity concerns for these monomers. (If data are submitted that show that no residual monomers are present, or that those residual monomers are entrapped in such a way that they cannot have an effect, then these data can be deleted.) The anhydride monomers (e.g., trimellitic anhydride) can be respiratory sensory irritants, and exposures in the workplace to the anhydrides that are used in the production of these copolymers have resulted in numerous adverse effects. Symptoms such as asthma, allergic rhinitis, bronchitis, conjunctivitis, rhinoconjunctivitis, and elevated antibody levels have been associated with occupational exposures (Table 2). The safety of several of the non-anhydride monomers as used in cosmetics has previously been reviewed by the Cosmetic Ingredient Review (CIR) Expert Panel (Panel). In 2012, the Panel concluded that adipic acid and sebacic acid are safe in the present practices of use and concentration,8 and that propylene glycol is safe as used in cosmetic formulations at the present practices of use and concentration when formulated to be non-irritating.26 In 1999, the Panel published a special report on the reproductive and developmental toxicity of ethylene glycol and its ethers, concluding that the metabolites of ethylene glycol monoalkyl ethers are reproductive and developmental toxins, but in general, these metabolites of concern are not expected to be formed in cosmetic formulations that contain polymers of ethylene glycol.11 CHEMISTRY Definition and Structure The ingredients in this report are related as copolymers that share in common trimellitic anhydride as a monomer. The monomers of these copolymers are interconnected via ester bonds to form highly branched polymeric (polyester) networks. For example, propylene glycol/sebacic acid/trimellitic anhydride copolymer is the result of the polymerization of propylene glycol, sebacic acid, and trimellitic anhydride. In this case, there are two acidic monomers (sebacic acid and trimellitic anhydride) and one alcohol monomer (propylene glycol). This means that to form a polyester copolymer, every other repeat unit must be propylene glycol. Whether the repeat unit on either end of propylene glycol is the residue of sebacic acid or the residue of trimellitic acid is dependent on the polymerization conditions. Since trimellitic anhydride serves as a ridged, nonlinear, tri-functional (trivalent) monomer, these polymers are branched, if not highly-branched in a manner similar to dendrimers.

Distributed for comment only -- do not cite or quote

Figure 1. propylene glycol/sebacic acid/trimellitic anhydride copolymer – the connectivity between alcohols (propylene glycol, in this case) and which acids (sebacic acid and trimellitic anhydride, in this case) can vary; this is just one theoretical example of a fragment within a polymeric network

The definitions and structures of the ingredients included in this review are provided in Table 1.

Physical and Chemical Properties Very little physical and chemical properties data on the trimellitic anhydride copolymers as cosmetic ingredients were found in the published literature, and unpublished data were not provided. The property information that was found is presented in Table 3.32-34 Method of Manufacture The manufacture of each of these trimellitic anhydride copolymers begins with trimellitic anhydride (an activated form of trimellitic acid). No ingredient-specific manufacturing flow charts or synthetic schemes have been submitted. However, general polyester synthetic techniques common in the art can be described.35 For example, synthesis would start by dissolv- ing trimellitic anhydride in a dry solvent, such as dimethylformamide, under an inert gas, such as nitrogen. Next, whatever polyol monomer (e.g., propylene glycol, neopentyl glycol, cyclohexanedimethanol, trimethylolpropane, ethylene glycol, or trimethylpentanediol) selected for use would be added to the solution and distilled under reduced pressure. In a second step, the resultant terminal alcohol groups may be further esterified with appropriate mono- or multi-functional carboxylic acids (e.g., sebacic acid, adipic acid, maleic anhydride, isostearic acid, phthalic anhydride, sebacic acid, or isophthalic acid), and likely a catalyst. This two-step methodology would result in dendrimer-like copolymers, with trimellitic acid as the central core in each. Alternatively, trimellitic anhydride, a polyol, and whatever additional carboxylic acid of choice could be reacted together in one step, under similar conditions, to form more random structured copolymers. Due to the trivalent nature of trimellitic acid, however, the resultant polymers from either of these two methodologies would be non-linear. Phthalic Anhydride/Trimellitic Anhydride/Glycols Copolymer Phthalic anhydride/trimellitic anhydride/glycols copolymer results from condensation of phthalic anhydride, trimellitic anhydride, ethylene glycol and neopentyl glycol monomers.36 Impurities/Constituents Very little impurities data on the trimellitic anhydride copolymers as cosmetic ingredients were found in the published literature, and unpublished data were not provided. The reactions used to manufacture the copolymers can be designed to result in little to no residual monomer, but information on the exact manufacturing process was not available. Without having a precise manufacturing method, the amount of monomer present in the copolymer is unknown. Therefore, it is possible that residual amounts of the following monomers (used in the production of these copolymers) could be present in these ingredients: adipic acid, cyclohexanedimethanol, ethylene glycol, isophthalic acid, maleic anhydride (or maleic acid), Distributed for comment only -- do not cite or quote

neopentyl glycol, phthalic anhydride (or phthalic acid), trimellitic anhydride (or trimellitic acid), trimethylpentanediol, trimethylolpropane, and isostearic acid. One supplier has reported that adipic acid/neopentyl glycol/trimellitic anhydride copolymer is sold in butyl acetate, and does not contain formaldehyde, toluene, or xylene.33 USE Cosmetic The safety of the cosmetic ingredients included in this assessment is evaluated on the basis of the expected use in cosmetics. The Panel utilizes data received from the U.S. Food and Drug Administration (FDA) and the cosmetics industry in determining the expected cosmetic use. The data received from the FDA are those it collects from manufacturers on the use of individual ingredients in cosmetics by cosmetic product category in its Voluntary Cosmetic Registration Program (VCRP). Those received from the cosmetic industry are submitted in response to a survey conducted by the Personal Care Products Council (Council) of the maximum reported use concentrations by product category. According to the 2015 VCRP data, adipic acid/neopentyl glycol/trimellitic anhydride copolymer is reported to be used in 411 cosmetic formulations, and phthalic anhydride/trimellitic anhydride/glycols copolymer is reported to be used in 74 cosmetic formulations, and both of these ingredients are reported to be used almost exclusively in nail formulations37 (Table 4). The results of the concentration of use survey conducted by the Council in 2015 indicate that the highest maximum concentration of use for both of these ingredients is in nail polish and enamel; in this product category, adipic acid/neopentyl glycol/trimellitic anhydride copolymer is reported to be used at up to 32.8%, and phthalic anhydride/trimellitic anhydride/ glycols copolymer is reported to be used at up to 12%.38 According to the Council survey, adipic acid/neopentyl glycol/trimellitic anhydride copolymer has a reported use in one product category that results in dermal exposure, i.e., it is used in face and neck products with a reported maximum concentration of use of 1%. None of the other trimellitic anhydride copolymers (i.e., adipic acid/CHDM/MA/neopentyl glycol/trimellitic anhydride copolymer; isostearoyl trimellitic anhydride/trimethylolpropane copolymer; propylene glycol/sebacic acid/trimellitic anhydride copolymer; or trimethylpentanediol/isophthalic acid/trimellitic anhydride copolymer) are reported to be in use. All of the trimellitic anhydride copolymers named in the report are not restricted from use in any way under the rules govern- ing cosmetic products in the European Union.39 Non-Cosmetic Adipic acid/CHDM/MA/neopentyl glycol/trimellitic anhydride copolymer is used in preparation of glass fiber reinforced plastic.32 Phthalic anhydride/trimellitic anhydride/glycols copolymer is used in the manufacture of dyes, pharmaceuticals, insecticides, and as a hardener for resins.40 TOXICOKINETICS Absorption, Distribution, Metabolism, and Excretion Toxicokinetics studies on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. TOXICOLOGICAL STUDIES Toxicological studies on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. REPRODUCTIVE AND DEVELOPMENTAL TOXICITY Reproductive and developmental toxicity data on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. GENOTOXICITY Genotoxicity data on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. CARCINOGENICITY Carcinogenicity data on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. Distributed for comment only -- do not cite or quote

IRRITATION AND SENSITIZATION Dermal Skin Irritation and Sensitization Dermal irritation and/or sensitization test data on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. However, several case reports describing allergic reactions to phthalic anhydride/trimellitic anhydride/ glycols copolymer were available.36,40,41 Details from these case reports are summarized in Table 5. Ocular Ocular irritation data on the trimellitic anhydride copolymers were not found in the published literature, nor were unpublished data provided. SUMMARY This report addresses the safety of 6 trimellitic anhydride copolymers as used in cosmetics. According to the International Cosmetic Ingredient Dictionary and Handbook, these ingredients are reported to function as film formers. The trimellitic anhydride copolymers are related as they all share a common monomer, i.e., trimellitic anhydride; each copolymer is also composed of another 1 to 4 monomers. The monomers that comprise these copolymers are interconnected via ester bonds to form highly branched polymeric (polyester) networks. Currently, no information is available regarding the amount of residual monomer in these copolymers. VCRP data obtained from the FDA, and data received in response to a survey of the maximum reported use concentration by product category conducted by the Council, indicate that 2 of the 6 ingredients included in this safety assessment are used in cosmetic formulations. Adipic acid/neopentyl glycol/trimellitic anhydride copolymer is reported to be used in 411 cosmetic formulations and phthalic anhydride/trimellitic anhydride/glycols copolymer is reported to be used in 74 cosmetic formulations; the highest maximum concentration of use for both of these ingredients is in nail polish and enamel, with maximum concentrations of use in this category reported as 32.8% adipic acid/neopentyl glycol/trimellitic anhydride copolymer and 12% phthalic anhydride/trimellitic anhydride/ glycols copolymer. Dermal irritation and/or sensitization test data on the trimellitic anhydride copolymers were neither found in the published literature, nor were unpublished data provided. However, several case reports described allergic reactions to phthalic anhydride/ trimellitic anhydride/ glycols copolymer. Toxicokinetics, toxicological, genotoxicity, or ocular irritation data were neither found in the published literature, nor were unpublished data provided. DISCUSSION To be developed. CONCLUSION To be determined.

Distributed for comment only -- do not cite or quote

TABLES Table 1. Definitions and Functions1 ,CIR Staff Ingredient (CAS No., if available) Definition Function(s) Adipic Acid/CHDM/MA/Neopentyl a copolymer of adipic acid, cyclohexanedimethanol (CHDM), maleic film former Glycol/Trimellitic Anhydride Copolymer anhydride (MA), neopentyl glycol and trimellitic anhydride monomers [67970-02-9] [The monomers are:

]

Adipic Acid/Neopentyl Glycol/Trimellitic a copolymer of adipic acid, neopentyl glycol and trimellitic anhydride film former Anhydride Copolymer monomers (28407-73-0) [The monomers are:

] Isostearoyl Trimellitic a copolymer of trimellitic anhydride and trimethylolpropane chain-terminated skin protectant; skin- Anhydride/Trimethylolpropane Copolymer by isostearic acid. conditioning agent - (1190965-82-2) emollient; skin- [The monomers are: conditioning agent - miscellaneous

] Distributed for comment only -- do not cite or quote

Table 1. Definitions and Functions1 ,CIR Staff Ingredient (CAS No., if available) Definition Function(s) Phthalic Anhydride/Trimellitic Anhydride/Glycols a copolymer of phthalic anhydride, trimellitic anhydride, ethylene glycol, and film former; viscosity Copolymer neopentyl glycol monomers increasing agent – (186688-25-5) non-aqueous [The monomers of are:

] Propylene Glycol/Sebacic Acid/Trimellitic a copolymer of propylene glycol, sebacic acid and trimellitic anhydride film former Anhydride Copolymer [The monomers are:

] Trimethylpentanediol/Isophthalic Acid/Trimellitic a copolymer of trimethylpentanediol, isophthalic acid and trimellitic anhydride film former; viscosity Anhydride Copolymer monomers increasing agent – non-aqueous [The monomers are:

]

Distributed for comment only -- do not cite or quote

Table 2. Monomer safety data Monomer (CAS No.) Summary Information Reference trimellitic anhydride molecular weight: 192.13 g/mol 2-7 (552-30-7) toxicokinetics: rapidly converted to acid trimellitic in the body (complete hydrolysis likely occurs in <10 min); in rats exposed to 0.95 mg/m3 for 45 min and killed 3 hr or 1, 2, 4, 8, 16, and 32 days following exposure, in general, the highest tissue concentrations were obtained at the first time point (T-max<3 hr), and a second T-max of 8 days was reported for lung lymph nodes in male rats, suggesting a possible role in the gender differences observed for lung toxicity; the biological half-life in the lungs was estimated to be 21 days in male rats and 16 days in female rats, and in lung associated lymph nodes, half-lives of 13 and 33 days were estimated for male and female rats, respectively; some of the rats became sensitized

oral toxicity: the oral LD 50 has been reported to range from 2,030 to 3,340 mg/kg in male and female rats, with stomach lesions appearing as the most consistent lesion upon necropsy; no adverse effects were observed in rats in a 90-day feed study with 1000- 10,000 ppm (50-500 mg/kg/day)

dermal toxicity: the dermal LD 50 was >2000 mg/kg in rabbits after a 24-h occlusive application 3 inhalation toxicity: the LC 50 in rats was reported to exceed a concentration of 2,330 mg/m (4-h exposure), with lung lesions appearing as the most consistent lesion upon necropsy numerous studies were conducted in rats, and in each of the following studies, the exposure was 6 h/day, 5 days/wk: in 2-wks studies, no adverse effects were observed with 0.3 mg/m3; in rats exposed to 0.1 mg/m3, lung injury was absent after 2 days of exposure, minimal after 6 days of exposure, and marked after 10 days of exposure; a dose-dependent increase in antibody levels and lung foci was observed in rats exposed to up to 0.30 mg/m3 for 1-2 wks, and the lung foci completely resolved within 12 days after the last exposure, but reappeared following exposure to a single challenge concentration; exposure to 0.5 mg/m3 produced hemorrhagic foci of the lung and increased antibody levels, and treatment with estrogen but not testosterone reduced the number of lung foci in both male and female rats; in a 13-wk study, a dose-dependent increase in lung lesions (hemorrhagic foci, inflammatory cell infiltration, bronchoalveoloar pneumonia) and antibody levels was observed in rats exposed to ≤0.054 mg/m3, and these effects were more pronounced in rats following 6.5 weeks of exposure than observed in animals following 13 wks of exposure, and a NOEL was not identified in 5-day inhalation study in mice with a 14-day recovery period, decreased time of inspiration and expiration and increased length of apneic periods were observed, and the LOEL was 0.01 mg/m3 reproductive and developmental toxicity: not teratogenic effects or fetal deaths in rats or guinea pigs exposed to 500 µg/m3 via inhalation on days 6-15 of gestation, however lung foci and TMA-specific antibody were observed in exposed dams and TMA-specific antibody was also noted in neonatal rats and in fetal but not neonatal guinea pigs, lung foci were only observed in the challenged offspring of rats whose mothers had not completely recovered from the original TMA exposure, but lung foci were not observed in adult rat offspring or in neonatal or adult guinea pig offspring; histopathological changes to reproductive tissues have not been observed in rats following subchronic exposures genotoxicity: not genotoxic with or without metabolic activation in 2 Ames tests ( ≤10,000 µg/plate), in a CHO/HGPRT mutation assay (≤2000µg/ml), or a chromosomal aberration assay in CHO cells (≤2080 µg/ml) dermal irritation/sensitizer: mild skin irritation potential; was a dermal sensitizer in guinea pigs with induction with a 30% solution in DMSO and challenge with 5% in acetone, but 300 mg of powder was not a sensitizer in guinea pigs; sensitizer in rats with 25-50% solutions in acetone/corn oil; sensitizer in mice with 10-50% solutions in acetone/olive oil ocular irritation: highly irritating to rabbit eyes when instilled undiluted effects with occupational exposure: may be a respiratory sensory irritant; elevated antibody levels, asthma, allergic rhinitis, and a late respiratory systemic syndrome are associated with occupational exposures in some workers immunologic response: TMA-induced syndromes are related to high chemical reactivity, which couples to human serum albumin and other proteins to form trimellitate protein conjugates, and some of the TMA syndromes have been correlated to immunologic responses to trimellitate haptenic groups; in TNF-α+/+ mice, in the late phase of TMA-induced contact hypersensitivity, the peak of ear swelling responses occurred at 24 h with single challenge and at 8 h after repeated challenge recommended limits: REL (U.S. NIOSH; exposure during a work week) – TWA 0.04 mg/m3 (0.005 ppm) adipic acid CIR Conclusion: safe in the present practices of use and concentration; reported to be used at a maximum 8 (124-04-9) of 0.000001% in leave-on formulations and 18% in rinse-off formulations molecular weight: 146.14 toxicokinetics: under normal physiological conditions, dicarboxylic acids are rapidly β-oxidized, resulting in very low cellular concentrations and practically non-detectable concentrations in the plasma, and oxidation of odd- and even-numbered chains proceeds to different end points with even chains completely, and odd-number chains not completely, oxidized; recovered unchanged in the urine of rats: with oral and i.v. dosing, approximately 53-67% and 59-71% was recovered, respectively; in humans, 6.76-61% was found in the urine after oral administration

oral toxicity: oral LD 50 in rats ranged from 0.94 g/kg to greater than the highest dose tested (11 g/kg); in feeding studies in rats, the NOAEL was >435 mg/kg bw/day in a 4 wk study; with ≤34000 mg/kg bw/day sodium adipate in a protein-deficient diet, the NOAEL was 3333 mg/kg bw in a 19-wk study; 10/14 rats fed a diet containing 3200 mg/kg bw/day died during wks 0-4 of a 33 wk study, body weights of the surviving animals in that group were similar to controls at study termination, and slight effects were seen in the liver; in a 2-yr study, the NOAEL was 1%; slight decreases in weight gains were observed with 3 and 5% inhalation toxicity: in mice exposed to 13 or 129 mg/m3 adipic acid via inhalation for 4 mos (details of exposure not provided), decreased weight gain, altered oxidase activity, and upper respiratory tract, liver, kidney, and central nervous system effects were observed Distributed for comment only -- do not cite or quote

Table 2. Monomer safety data Monomer (CAS No.) Summary Information Reference reproductive and developmental toxicity: NOAELs for maternal and developmental toxicity were 288 mg/kg bw in CD-1 mice (highest dose given orally on days 6-15 of gestation) and 288 mg/kg bw in Wistar rats (highest dose given orally on days 6-15 of gestation); NOAELs for maternal toxicity and developmental toxicity were ≥250 mg/kg bw and 250 mg/kg bw, respectively, in Dutch-belted rabbits (maximum dose given by gavage on days 6- 18 of gestation) genotoxicity: not genotoxic in vitro in the Ames tests (concentrations up to 10,000 mg/plate), yeast gene mutation assay (concentrations ≤200 mg/l), mouse lymphoma assay (concentrations of ≤2000 µg/plate), or cytogenetic assay (human embryonic lung fibroblast cells, ≤200 mg/l), or in vivo in a cytogenetics assay (chromosomes from rats dosed by gavage with 5000 mg/kg bw (single dose) or 2500 mg/kg bw (1x/day for 5 days)) or in a dominant lethal assay (up to 5000 mg/kg bw); a significant increase in resorption per implant site was observed in hamsters with 205 mg/kg bw adipic acid, resulting in a decreased number of live fetuses (this decrease was not evaluated statistically, and no effects were reported at this or the other doses (≤44 mg/kg bw)) carcinogenicity: not carcinogenic in a 2-yr chronic study in rats fed up to 5% adipic acid dermal irritation: occlusive application of a 50% aq. solution to intact and abraded rabbit skin for 24 h produced an erythema score of 2-3/4 for intact skin, with clearing by day 3, and mild to severe erythema and edema, for abraded skin, which cleared by day 7; when applied undiluted or as an 80% aq. paste to the backs or ears of rabbits for 24 h, no irritation was observed on the backs and erythema was observed on the ear, with clearing by 72 h; no irritation was observed with a 24 h application (not details provided; semi-occlusive application of a 50% paste in propylene glycol produced slight to mild irritation in 3/6 rabbits; a semi-occlusive application of undiluted adipic acid was not corrosive; 50% in propylene glycol, was not irritating to guinea pigs dermal sensitization: not a sensitizer (1% for intradermal induction, up to 50% in propylene glycol for derma challenge) and produced very mild or no irritation ocular irritation: mild to severe ocular irritant in rabbit eyes; irritation was dose-dependent peroxisome proliferation: did not induce peroxisome proliferation and did not affect relative liver to body weights in rats fed 2% dissolved in alcohol for 3 wks cyclohexanedimethanol molecular weight: 144.21 g/mol 9,10 (105-08-8) toxicokinetics: when rats were dosed by gavage with 40 or 400 mg/kg/bw (14C; 70% trans-, 30% cis-isomers), there was rapid absorption from the GI tract and after 48 hr, 95% of dose was excreted in urine, 2.5% in feces, 14 0.03% respired as CO2, and 0.4% remained in carcass, recovery of radioactivity averaged 98.9% of dose, and the half-life in plasma from rats dosed with 400 mg/kg was about 13 min; the test article and 4-hydroxymethyl- cyclohexanecarboxylic acid were detected in plasma, unchanged test article was not detected in urine and the major metabolites identified in urine were cyclohexanedicarboxylic acid (68%) and 4-hydroxymethylcyclo- hexanecarboxylic acid (31%), <2% of radioactivity in urine was not fully characterized, and the cis-trans ratio of metabolites excreted in urine was the same as that of original dose oral toxicity: observations in rats given a single oral dose by gavage of 400-6400 mg/kg bw ranged from normal to very weak with prostration and vasodilatation; the oral LD 50 is reported to be >3200-6400 mg/kg in rats; in a 13-wk study in which rats were given 4-12.5 mg/l (i.e., 256-861 mg./kg bw/day for males and 440-1754 mg/kg bw/day for females) in drinking water, the LOAELs were 861 and 1754 mg/kg bw/day for males and females, respectively, based on decreased survival, abnormal urine and feces, reduced body weights and weight gains, decreased feed consumption and increased urinary protein levels, and the NOAELs were 479 and 754 mg/kg bw/day for males and females, respectively reproductive and developmental toxicity: in male and female rats were given 4-12.5 mg/l (i.e., 256-861 mg./kg bw/day for males and 440-1754 mg/kg bw/day for females) in drinking water prior to and during mating and until day 4 of lactation, for systemic toxicity, the LOAELs were 861 and 1360 mg/kg bw/day (based on bloody or discolored urine, reductions in body weights and weight gains and decreased feed consumption) and the NOAELs were 479 and 854 mg/kg bw/day for males and females, respectively, for developmental toxicity the LOAEL was 1360 mg/kg bw/day (based on decreased fetal weight and postnatal survival) and the NOAEL was 854 mg/kg bw/day, and for reproductive toxicity the LOAEL was greater than the highest dose tested and the NOAEL was 1360 mg/kg bw/day genotoxicity: not genotoxic in vitro in an Ames test at ≤5000 µg/plate without and with activation, in a chromosomal aberration assay in human lymphocytes at ≤10 mM with or without metabolic activation, or in vivo at doses up to 2000 mg/kg in a mammalian bone marrow chromosomal aberration test in rats or a micronucleus test in mice dermal irritation/sensitization: not irritating to rat skin when applied undiluted with a semi-occlusive patch; not a sensitizer in a guinea pig maximization test with 1% at intradermal induction at neat, applied neat for topical induction, and neat and at 50% for challenge ocular irritation: moderately irritating to rabbit eyes ethylene glycol CIR Conclusion (Special Report on reproductive and developmental toxicity): the metabolites of ethylene 11 (107-21-1) glycol monoalkyl ethers are reproductive and developmental toxins, but in general, these metabolites of concern are not expected to be formed in cosmetic formulations that contain polymers of ethylene glycol molecular weight: 62.07 g/mol Distributed for comment only -- do not cite or quote

Table 2. Monomer safety data Monomer (CAS No.) Summary Information Reference isophthalic acid molecular weight: 166.13 g/mol 12-14 (121-91-5) toxicokinetics: readily eliminated from the body, largely unchanged, thru the urine; in a 13-wk feed study in rats, levels of test article in the blood increased in a dose-dependent manner, and 24-h urinary excretion samples collected on days 7, 30, 60, 90 indicate that urinary excretion, presumably as the unchanged chemical, is the primary route of excretion.; exposure of rats to 10 mg/m3 for 6 h/day resulted in immediate detection of the test article in the blood, serum levels were 5.3-9.3 ug/ml in females and 1.4-3.4 ug/ml for males, but no test article was detected in the blood 1 wk following exposure

oral toxicity: reported oral LD 50s have been reported in the range >5000 to 13,000 mg/kg in rats; in a 13-wk feeding study in rats with 800 mg/kg/day, a slight increase in the incidence of and renal pathology (mild hydronephrosis, pelvic calcification) were observed, the NOAEL was 250 mg/kg/day, and the LOAEL was 800 mg/kg/day dermal toxicity: no mortality was observed in rabbits following acute exposure to 23,000 mg/kg; in a study with a 24 h occlusive application, the oral LD50 was >2000 mg/kg bw inhalation toxicity: no mortality was observed in rats following acute inhalation exposures to 11,400 mg/m3 ; in a 4-wk study in which rats were exposed 6h/day 5 days/wk, no significant effects were observed with up to 10 mg/m3, and the NOAEL was 10 mg/m3 reproductive and developmental toxicity: no maternal or developmental toxicity at inhalation exposures up to 10 mg/m3 in rats on days 6-15 of gestation genotoxicity: mixed results were reported in 3 Ames tests; negative in a chromosomal aberration assay in CHO cells at up to 5000 µg/ml with and without metabolic activation, in an HGPRT assay, and a mouse lymphoma mutation assay (with metabolic activation) dermal irritation/sensitization: no dermal irritation with application of a single dermal dose to rabbits; a 4-h semi- occlusive application and a 24-h occlusive application of undiluted test material was not irritating to rabbit skin; mild dermal irritation with application of 2000 mg/kg; a 30% solution was not a skin sensitizer in guinea pigs (a reaction were observed in 1/10 of the guinea pigs) ocular irritation: 0.1 g, undiluted, was non- to slightly irritating to rabbit eyes maleic anhydride toxicokinetics: readily hydrolyzed to maleic acid under aqueous conditions 15-18

(108-31-6) oral toxicity: according to the OECD summary, relatively low acute toxicity, with the oral LD50 of about 1.0 g/kg in rats, but according to the EPA summary, acute oral studies in rats, mice, rabbits, and guinea pigs have demonstrated moderate to high acute toxicity by ingestion; oral feeding studies have resulted in kidney damage in rats at relatively high doses (> 100 mg/kg/day after 90 days of exposure), with the effects (which were more severe in males than in females) likely due to maleic acid; no kidney effects were observed in rats that were fed diets containing 32 and 100 mg/kg/day maleic anhydride for 2 yrs; a dietary study in dogs dosed with up to 60 mg/kg for 7days/wk for 90 days, showed no adverse effects related to maleic anhydride exposure

dermal toxicity: according to the OECD summary, relatively low acute toxicity, with a dermal LD50 in the range of 1.6 - 2.6 g/kg in rabbit, but according to the EPA summary, moderate acute dermal toxicity inhalation toxicity: bronchial asthma was observed with acute exposure in guinea pigs; possible respiratory sensitizer to rats; repeated exposure by inhalation to rats, hamsters, and monkeys resulted in effects that were limited to the respiratory tract and eye irritation; in a 4-wk study in rats exposed 6 hours/day to up to 84 mg/m3 (21 ppm), evidence of nasal, trachea, and lung irritation was observed at all exposure levels, and the effects were concentration-related and included epithelial hyperplasia and the presence of inflammatory exudates in the nasal turbinates and trachea; and epithelia hyperplasia, squamous metaplasia, and intra-alveolar hemorrhage in the lung, with a LOAEL of 12 mg/m3 (3 ppm); in a 6-month inhalation study in which rats, hamsters, and monkeys were exposed to up to 9.8 mg/m3 (2.4 ppm), respiratory tract and eye irritation were observed, hyperplastic and metaplastic changes in the nasal passages were considered indicative of irritation and judged to be reversible, and the NOAEL for rats was 3.3 mg/m3 (0.8 ppm) and the NOAEL for hamsters and monkeys was 9.8 mg/m3 (2.4 ppm) reproductive and developmental toxicity: in a 2-generation reproductive toxicity study in which rats were dosed via gavage with up to 150 mg/kg/day, the NOAEL for reproductive effects was 55 mg/kg/day (highest dose tested due to parental death at 150 mg/kg/day), but in the parental group adverse effects (mortality, body weight changes, and respiratory irritation) were observed at 150 mg/kg/day and there were histopathological effects in the kidneys and bladder of the parental animals (first generation only) in all treated dose groups, and the LOAEL for parental effects was 20 mg/kg/day; no developmental toxicity was observed when pregnant rats were dosed via gavage with up to 140 mg/kg/day, but the dams in all dose groups either lost weight or failed to gain weight between days 6 and 9 of gestation (effect was not statistically significant at any interval and was reversible), and the NOAEL (maternal) was determined to be 140 mg/kg/day genotoxicity: negative in bacterial gene mutation tests; in a single in vitro chromosomal aberration test with and without S-9 was positive (due to inadequate documentation on this study, it is unclear whether the results were due to the test material itself or a change in pH to an acidic environment, which could have resulted in a nonspecific effect); exposure of Sprague-Dawley rats of up to 100 mg/m3 (25 ppm) did not increase chromosomal aberrations in the bone marrow carcinogenicity: not carcinogenic when given to rats in their diets for 2 yrs at doses up to 100 mg/kg/day; the EPA has not classified maleic anhydride for carcinogenicity dermal irritation/sensitization: severely irritating to the skin of rabbits after application of 0.5 g to the skin for 4 hours (mean irritation score of 3.67 – 4.00 during the 7-day observation period); shown to be a skin sensitizer to guinea pigs ocular irritation: severely irritating to the eyes of rabbits (eye irritation score was 106.7/110); in humans, Distributed for comment only -- do not cite or quote

Table 2. Monomer safety data Monomer (CAS No.) Summary Information Reference exposure to 0.25 - 0.38 ppm (1- 1.6 mg/m3) produced eye irritation, no irritation was reported at 0.22 ppm respiratory effects (human): exposure to 0.25 - 0.38 ppm (1- 1.6 mg/m3) produced respiratory tract irritation, no irritation was reported at 0.22 ppm effects with occupational exposure: according to the OECD summary, there have been a few published human cases suggesting that maleic anhydride provokes asthma in a relatively small proportion of exposed workers, however, questions have been raised whether the asthma was actually related to maleic anhydride exposure; according to the EPA summary, chronic (long-term) exposure has been observed to cause chronic bronchitis, asthma-like attacks, and upper respiratory tract and eye irritation in workers, and in some people, allergies have developed so that lower concentrations can no longer be tolerated recommended limits; the EPA RfD is 0.1 mg/kg bw/d based on renal lesions in rats (the RfD is an estimate (with uncertainty spanning perhaps an order of magnitude) of a daily oral exposure to the human population (including sensitive subgroups) that is likely to be without appreciable risk of deleterious noncancer effects during a lifetime; it is not a direct estimator of risk but rather a reference point to gauge the potential effects); OEHHA inhalation REL (chronic exposure level, exposure 24 h/day for a lifetime) is 0.7 µg/m3 (2.5 ppb) Maleic Acid - CIR Conclusion: safe for use in cosmetic formulations as a pH adjustor in the practices of 19 use described in the safety assessment (only use as a pH adjustor was evaluated) ; was reported to be used at 0.004% in “other” bath products; used in hair straighteners, other hair-coloring preparations, and shaving cream according to VCRP data, but no concentration data were reported for these uses according to the Discussion of the report, the Panel recognized that maleic acid itself may be a dermal and/or ocular irritant, but its use as a pH adjustor in cosmetic formulations dictates that most of the acid would be neutralized into various maleate salts; additionally, the concentration of maleic acid would be dependent on the alkaline content of the formulation; the Panel also stated that safety of maleic acid as a pH adjustor should be based on the amount of free maleic acid remaining after neutralizing the formulations neopentyl glycol molecular weight: 104.15 20,21 (126-30-7) toxicokinetics: after rabbits were dosed with 1-1.5 g/kg bw (unlabeled) by gavage, 62% of the dose was found in the 24-h urine as the conjugate of glucuronic acid, indicating rapid absorption, 1.9% was recovered as the metabolite 3-hydroxy-2,2-dimethylpropionic acid, and only 0.7% of the dose was present unchanged

oral toxicity: oral LD 50 in rats reported as 3200 mg/kg to 6920 mg/kg; with repeated dosing with up to 1000 mg/kg/day by gavage, the NOAEL for males and females were 300 and 1000 mg/kg/day, respectively; in a 90- day gavage study, no adverse effects were observed at doses up to 1000 mg/kg

dermal toxicity: the dermal LD50 was >4000 mg/kg bw in guinea pigs with a 24-h occlusive application to a 20% solution in acetone/corn oil vehicle inhalation toxicity: there was no mortality in rats exposed to a saturated vapor for 8 h (the calculated nominal concentration was 140 mg/m³); rats exposed for 6 h to 39,400 ppm (168 mg/l) showed symptoms of irritation of the respiratory tract, 1/3 died within 24 h; 3 rats were exposed to 4000 ppm, 6 h/day for 10 days, and irritation of the respiratory tract and dilatation of the skin blood vessels were detected, but no evidence f toxic effects on internal organs was observed at necropsy reproductive and developmental toxicity: the NOAEL f or maternal and prenatal developmental toxicity was 1000 mg/kg with dosing with ≤1000 mg/kg on days 6-19 of gestation; in a reproductive toxicity study in which rats were dosed with up to 1000 mg/kg/day by gavage before and during mating and through day 3 of lactation, the NOAEL was 1000 mg/kg for the parent and F1 generation genotoxicity: not genotoxic in an Ames test (up to 5000 µg/plate) or to Chinese hamster ovary cells (up to 1 mg/ml) dermal irritation/sensitization: an 80% aq. solution was not an irritant to rabbit skin with a 20-h -occlusive exposure; slightly irritating to rabbit skin with a 4-h exposure to undiluted test material; not a sensitizer in a mouse LLNA (60% in propylene glycol) ocular irritation: instillation of crystalline test substance resulted in serious damage to rabbit eyes in one study, and instillation of neat test article in rabbits resulted in irreversible damage in one study and slight to moderate irritation in another study phthalic anhydride molecular weight: 148.12 g/mol 22-25 (85-44-9) stability: rapidly hydrolyzed to phthalic acid upon contact with water toxicokinetics: half-life was 30.5 sec at pH 7.24. and 61 sec at pH 6.8; unconjugated compound was found in the urine of humans exposed by inhalation, demonstrating systemic absorption and elimination via the urine and the existence as a hydrolysis product in vivo

oral toxicity: oral LD 50 of 1530 mg/kg bw in rats; low repeated dose toxicity in rats; in a 7-wk feed study in which mice were given up to 7140 and rats up to 3330 mg/kg bw/day, there were no effects in mice, but there was a significant reduction in body weight gains in rats of the high dose group and centrilobular cytoplasmic vacuolation were seen in the livers in most of the male rats fed 1660 mg/kg bw/day; in male and female mice fed 4670 and 3430 mg/kg bw/day, respectively, for 105 wks, significantly increased incidences of lung and kidney lymphocytosis in the low- and high-dose males and females, chronic bile duct inflammation in the high-dose males and females, and dose-related adrenal atrophy and mineralization of the thalamus in the low-and high-dose males were observed, and the time weighted LOAELs were approximately 1717and 2340 mg/kg bw/day in female and male mice, respectively, and no NOAEL was obtained

dermal toxicity: the dermal LD50 was >10,000 mg/kg bw in rabbits

inhalation toxicity: the LC 50 was >2.14 mg/l following a 4-h nose-only exposure; respiratory sensitization Distributed for comment only -- do not cite or quote

Table 2. Monomer safety data Monomer (CAS No.) Summary Information Reference potential in guinea pigs, and animals exposed to and challenged with 5.0 mg/m3 phthalic anhydride dust had significant numbers of hemorrhagic lung foci reproductive and developmental toxicity: no evidence of toxicity to reproductive organs was observed in comprehensive carcinogenicity studies in rats and mice, as no treatment-related changes were observed for any reproductive organ investigated during macroscopic and microscopic examination; decreased spermatozoa motility time was reported in one study in which male rats were exposed via inhalation genotoxicity: not mutagenic in the Ames test (up to 5000 µg/plate); induced chromosomal aberrations in mammalian cells only at the high concentration of 10 mM without metabolic activation; not genotoxic in a sister chromatid exchange assay (10-300 µg/ml with and without metabolic activation) carcinogenicity: no evidence of carcinogenicity in rats fed 1000 mg/kg bw/day or in in male and female mice fed 4670 and 3430 mg/kg bw/day, respectively, 105 wks dermal irritation/sensitization (animal): mild irritation was observed when 550 mg of flakes were applied to rabbits for 4 h, the average dermal irritation index after semi-occlusive exposure was 1.21/8, and observed effects were reversible; semi-occlusive application of 500 mg (moistened with water) on the ear for 24 h was not an irritant; sensitizer in the GPMT and at a concentration of 10% in the LLNA in mice dermal irritation (humans): contact with either the solid or a vapor has been reported to cause skin irritation after occupational exposure; erythema, blistering, ulceration and necrosis have been reported following skin contact ocular irritation: irritating to rabbit eyes (undiluted) mucous membrane irritation: a primary irritant in humans when in the form of vapor, fumes, or dust respiratory effects (humans): a primary irritant to the upper respiratory tract when in the form of vapor, fumes, or dust effects with occupational exposure: conjunctivitis, rhinitis, rhinoconjunctivitis, bronchitis, and irritation of the skin and mucous membranes of the respiratory tract have been observed in exposed workers; initial exposure produces coughing, sneezing, burning sensations in the nose and throat, and increased mucous secretion; occasional bloody sputum, emphysema, lower blood pressure, and minor signs of central nervous system excitation has been observed in chronically-exposed workers recommended limits: the EPA has calculated a provisional RfC of 0.12 mg/m3 (the RfC is an estimate (with uncertainty spanning perhaps an order of magnitude) of a continuous inhalation exposure to the human population (including sensitive subgroups) that is likely to be without appreciable risk of deleterious noncancer effects during a lifetime. It is not a direct estimator of risk but rather a reference point to gauge the potential effects); the EPA has established a RfD of 2.0 mg/kg bw/day propylene glycol CIR Conclusion: safe in the present practices of use and concentration when formulated to be non- 26,27 (57-55-6) irritating; reported to be used at up to 73% in leave-on formulations molecular weight: 76.09 g/mol toxicokinetics: the major route of metabolism is to lactaldehyde and then lactate via hepatic alcohol and aldehyde dehydrogenases dermal absorption: penetration from a ternary cosolvent solution through hairless mouse skin was 57% over a 24- h period, and it appears that it does not reach the dermis; it can act as a penetration enhancer

oral toxicity: the oral LD 50 was 21 g/kg in rats; no mortality in mice given 10% in drinking water for 14 days; lesions were not observed in rats that were fed diets containing 50,000 ppm (2.5 g/kg/day) for 15 wks or up to 50,000 ppm in the diet for 2 yrs; similar results were reported in a study in which dogs were fed 2 or 5 g/kg in the diet for 103 wks; in dogs given 5% in drinking water for 5-9 mos, no hepatic or renal impairment was noted inhalation toxicity: some effects in rats due to exposure of 2.2 mg/L air for 6 h/day, 5 days/wk, for 13 wks, but these effects were inconsistent and without dose-response trends reproductive and developmental toxicity: not teratogenic in female CD-1 mice when administered at a concentration of 10,000 ppm on days 8-12 of gestation; no reproductive effects in a continuous breeding reproduction study in albino mice with up to 5% administered in feed or water; no adverse effects on reproduction or development when given orally at at doses up to 1600 mg/kg in mice and rats, 1230 mg/kg in rabbits, or 1550 mg/kg in hamsters genotoxicity: no mutagenic in an Ames test at up to 10,000 µg/plate with or without metabolic activation; caused a dose-dependent increase in the frequency of SCEs in a Chinese hamster cell line and was classified as a weak inducer of SCEs, but in another SCE study, it was not genotoxic in human cultured fibroblasts or a cultured Chinese hamster cell line with or without metabolic activation; chromosomal aberrations were induced in Chinese hamster fibroblasts in another assay; was not genotoxic in additional tests for chromosomal aberrations, mitotic recombination, or basepair substitution, or DNA damage or a micronucleus test carcinogenicity: not carcinogenic in a 2-yr feeding study at up to 50,000 ppm in rats; did not induce skin tumors and was not carcinogenic in a lifetime dermal study at concentrations up to 100% in mice dermal irritation/sensitization (animal): 10% was not a dermal irritant in a 24-h test in nude mice, but hypertrophy, dermal inflammation, and proliferation were observed at a concentration of 50%; not an irritant to intact or abraded skin of rabbits in a Draize test, to guinea pig or rabbit skin when applied for 48 h using open and occlusive patches, or to swine skin in 48-h and 21-day open and occlusive patch tests; not a sensitizer in a mouse external ear swelling sensitization test undiluted), in a GPMT, OET, or chamber (Finn chamber) test (at 70%); was a potentially weak sensitizer in one maximization test, but the results of 6 other guinea pig sensitization indicated it was not an allergen dermal irritation/sensitization (human): induced skin irritation and sensitization reactions in normal subjects and patients in studies with test concentrations ranged from 2 - 100%, and reactions were observed at concentrations Distributed for comment only -- do not cite or quote

Table 2. Monomer safety data Monomer (CAS No.) Summary Information Reference as low as 10% in predictive tests and as low as 2% in provocative tests; however, the dermal irritation potentials of deodorant formulations containing 68.06% or 69.15% were evaluated in an SIOPT and compared to a reference in-use control formulation, and the formulations containing propylene glycol were no more irritating or even less irritating than the reference control; use studies of deodorant formulations containing 35%-73% did not report any potential for eliciting irritation or sensitization; deodorant formulations containing 69.15% or 86% did not induce sensitization reactions; questionable results were obtained in an RIPT of a deodorant containing 73%; retrospective analysis of pools of patient patch test data indicated that ≤6.0% of patients tested had positive reactions to a 30% aq. solution; increased the allergic responses in 43 patients patch tested with 50 µg 1% nickel sulfate solution photoallergenicity: did not produce a photoallergic response in a provocative photopatch test (concentration not specified) ocular irritation: did not induce corneal damage in rabbit eyes in one study; was a slight ocular irritant in another sebacic acid CIR Conclusion: safe in the present practices of use and concentration; reported to be used at a maximum 8 (111-20-6) of 0.03% in leave-on formulations and 1% in rinse-off formulations molecular weight: 202.25 toxicokinetics: oxidized to water and carbon dioxide, passing through acetyl-CoA and succinyl-CoA formation; under normal physiological conditions, dicarboxylic acids are rapidly β-oxidized, resulting in very low cellular concentrations and practically non-detectable concentrations in the plasma, and oxidation of odd- and even- numbered chains proceeds to different end points with even chains completely, and odd-number chains not completely, oxidized; recovered unchanged in the urine of rats: with i.v. dosing, approximately 35% sebacate was recovered oral toxicity: . disodium sebacate was not toxic to rats or rabbits fed up to 1000 mg/kg bw for 6 mos reproductive and developmental toxicity: in dietary studies, disodium sebacate was not a developmental toxicant in rats (500 mg/kg bw day; days of dosing not stated) or rabbits (1000 mg/kg bw) trimethylpentanediol molecular weight: 146.23 28,29

(144-19-4) oral toxicity: oral LD50 was 1800 mg/kg bw and >2000 mg/kg bw in male mice and female rats, respectively; in rats fed ≤2% for 57 days, the NOAEL for males and females was 0.5% (376 mg/kg bw/day) because of changes in organ wts at higher doses

dermal toxicity: in a study in guinea pigs (1 animal/dose), the dermal LD50 was >1 g/kg bw

inhalation toxicity: the LC 50 was >4.5 mg/l in rats with a 6-h exposure period (the respirable concentration of the test substance was 1.43 mg/l and the calculated exposure to the respirable portion of the test substance was 2.8 g/kg bw) reproductive and developmental toxicity: in a 3-generation dietary study in rats at concentrations up to 1% in feed, the NOAEL was 1% for both parental and for reproductive and developmental toxicity genotoxicity: not genotoxic, with or without metabolic activation, in an Ames test at concentrations up to 5000 µg/plate, in a mammalian cell assay at concentrations up to 1462.3 µg/mL in mouse lymphoma cells, or in a chromosomal aberration assay at concentrations up to 1500 µg/ml in CHO cells irritation/sensitization: an occlusive 4-h exposure was not irritating to intact or abraded rabbit skin; 0.1 M was not a sensitizer in guinea pigs ocular irritation: undiluted test material was a slight irritant to rabbit eyes trimethylolpropane molecular weight: 134.17 g/mol 30,31

(77-99-6) oral toxicity: in a gavage study, the LD50 was ~14,700 mg/kg in male rats; in a 90 day feeding study with up to 1.0% (667 mg/kg bw/d) in rats, the NOAEL was 0.1% (67 mg/kg bw/d) based on significant changes of clinical chemistry or hematological data (at ≥0.3 %) and histopathological changes, mainly in liver and spleen (at 1%)

dermal toxicity: the dermal LD50 was >10,000 mg/kg bw in rabbits with a 24-h occlusive application 3 inhalation toxicity: the LC 50 was >850 mg/m (vehicle, ethanol/lutrol 1:1) in rats with a 4-h exposure period reproductive and developmental toxicity: in rats dosed by gavage with up to 1000 mg/kg/day on days 6-20 of gestation, the NOAEL was 100 mg/kg bw/day for developmental toxicity due to decreases in fetal weight and 100 mg/kg bw/day for maternal toxicity due to decreased body weight and weight gains; in a study in which male and female rats were dosed by gavage with up to 800 mg/kg bw/day before, during, and after mating, the NOAEL was 800 mg/kg bw/day for reproductive and developmental toxicity and 200 mg/kg bw/day for general toxicity genotoxicity: not genotoxic, with or without metabolic activation, in an Ames test at concentrations up to 5000 µg/plate, in a mammalian cell assay at concentrations up to 1400 µg/mL in Chinese hamster lung fibroblasts, or in a chromosomal aberration assay at concentrations up to 1340 µg/ml in Chinese hamster lung cells dermal irritation/sensitization: application to the inner ear of rabbits did not produce irritation; not sensitization in a mouse LLNA at concentrations up to 50% ocular irritation: not irritating to rabbit eyes when instilled undiluted or melted Abbreviations: DMSO – dimethyl sulfoxide; EPA – Environmental Protection Agency; CHO – Chinese hamster ovary; GPMT – guinea pig maximization test; HGPRT - hypoxanthine-guanine phosphoribosyltransferase; LLNA – local lymph node assay; LOAEL – lowest observable adverse effect level; NIOSH – National Institute for Occupational Safety and Health; NOAEL – no-observable adverse effect level; OECD – Organisation for Economic Co-operation and Development; OEHHA - Office of Environmental Health Hazard Assessment; OET – open epicutaneous test; NIOSH REL – recommended exposure limit; OEHHA REL – reference exposure level; RfC – reference concentration; RfD - reference dose; RIPT – repeated insult patch test; SCE – sister chromatid exchange; SIOPT – single insult occlusive patch test; TMA – trimellitic anhydride; TWA – time-weighted average

Distributed for comment only -- do not cite or quote

Table 3. Physical and chemical properties Property Description Reference Adipic Acid/CHDM/MA/Neopentyl Glycol/ Trimellitic Anhydride Copolymer physical characteristics liquid or solid 32 solubility insoluble in water 32 Adipic Acid/Neopentyl Glycol/Trimellitic Anhydride Copolymer physical characteristics clear liquid 33 molecular weight 442.41g/mol 34 solubility low solubility in cold water; soluble in all “common solvents” 33 acid number 10-20 mg KOH/g 33

Table 4. Frequency and concentration of use according to duration and type of exposure # of Uses37 Max Conc of Use (%)38 # of Uses37 Max Conc of Use (%)38 # of Uses Max Conc of Use (%) Adipic Acid/Neopentyl Glycol/Trimellitic Phthalic Anhydride/ Trimellitic Anhydride Copolymer Anhydride/Glycols Copolymer Totals* 411 1-32.8 74 1.8-12 Duration of Use Leave-On 410 1-32.8 74 1.8-12 Rinse-Off 1 NR NR NR Diluted for (Bath) Use NR NR NR NR Exposure Type Eye Area NR NR NR NR Incidental Ingestion NR NR NR NR Incidental Inhalation-Spray NR NR NR NR Incidental Inhalation-Powder NR NR NR NR Dermal Contact 1 1 2 NE Deodorant (underarm) NR NR NR NR Hair - Non-Coloring NR NR NR NR Hair-Coloring NR NR NR NR Nail 410 5.4-32.8 72 1.8-12 Mucous Membrane NR NR NR NR Baby Products NR NR NR NR

* Because each ingredient may be used in cosmetics with multiple exposure types, the sum of all exposure types may or may not equal the sum of total uses.

Distributed for comment only -- do not cite or quote

Table 5. Case reports Case History Patch Testing Reference Phthalic Anhydride/Trimellitic Anhydride/Glycols Copolymer a female subject had a 2-mo history of an intermittent itchy - Patch testing was performed with the British standard series, a modified cosmetic 41 rash on the neck and around the eyes, associated with series, a modified plant series, with her cosmetics, and her nail varnish (which was episodes of swelling of the eyelids; mild erythema and applied to a Finn Chamber and allowed to dry before application); a positive scaling were present the upper eyelids and neck reaction (+) was observed with the nail varnish at days 2 and 4; there was no reaction to tosylamide/formaldehyde resin - Patch testing with the nail varnish components resulted in a positive reaction (++) to 1% phthalic anhydride/trimellitic anhydride/glycols copolymer in petrolatum was observed at days 2 and 4; a + reaction was also observed to the 5 coloring bases that contained this ingredient - testing in 12 control subjects had negative results a female subject had a 12-mo history of periorbital and - the subject was patch-tested with the BCDSB, a cosmetic and medicament series, 40 fingertip eczema and her own cosmetics; a positive reaction (+) was observed with her nail varnish (which was applied to a Finn Chamber and allowed to dry before application) on day 4 - Patch testing with the nail varnish components resulted in a positive reaction (++) to 1% phthalic anhydride/trimellitic anhydride/glycols copolymer in petrolatum at day 4 a female subject had a 6-mo history of perioral eczema and - the subject was patch-tested with the BCDSB, a cosmetic and dental series, and 40 dry, fissured lips her own cosmetics; a positive reaction (+) was observed with her nail varnish (which was applied to a Finn Chamber and allowed to dry before application) on day 4 a female subject had a 6-mo history of intermittent eczema - the subject was patch-tested with the BCDSB, a cosmetic series, and her own 40 of her face and fingers cosmetics; positive reactions (+) were observed with her nail varnish (which was applied to a Finn Chamber and allowed to dry before application) on days 2 and 4 - patch testing with the nail varnish components resulted in a positive reaction (+) to 1% phthalic anhydride/trimellitic anhydride/glycols copolymer in butyl acetate on day 4 a female subject had periungual dermatitis affecting all her - the subject was patch-tested with the BCDSB and acrylic series; she had positive 40 fingers; she had been wearing nail varnish and acrylic nails reactions to several compounds, including 1% phthalic anhydride/trimellitic for several years anhydride/glycols copolymer in butyl acetate on day 4on days 2 and 4 Three case reports were described: - patch testing with all nail polish formulations resulted in positive reactions to a 36 an atopic female subject presented with generalized few of the polishes; patch testing with toluene sulfonamide formaldehyde resin was eczema and major lichenification of lips and eyelids negative progressively worsening during the past 4 yrs; recent patch - patch testing with the nail varnish components identified phthalic anhydride/tri- testing with the standard ICDRG series (2) was negative mellitic anhydride/glycols copolymer (tested at 1% and 5% in petrolatum) as the allergen in all 3 subjects a female subject had recurrent eyelid and neck dermatitis, - subsequent patch testing with the raw monomers, i.e. phthalic anhydride, with severe nail dystrophy; patch testing with a standard trimellitic anhydride and glycols was performed in 2 subjects; the results were series, toluene sulfonamide formaldehyde resin, and her negative own nail polishes were negative; the lesions persisted for several mos, and she was referred again a female subject had a history of occupational contact dermatitis and asthma due to glutaraldehyde presented with head and neck eczema lasting for 3 mos

Abbreviations: BCDSB - British Contact Dermatitis Society Standard Battery; ICDRG - International Contact Dermatitis Research Group

Distributed for comment only -- do not cite or quote

REFERENCES

1. Nikitakis J and Breslawec HP. International Cosmetic Ingredient Dictionary and Handbook. 15 ed. Washington, DC: Personal Care Products Council, 2014.

2. European Chemical Agency (ECHA). Benzene-1,2,4-tricarboxylic acid 1,2-anhydride (trimellitic anhydride; CAS No. 552-30-7). http://apps.echa.europa.eu/registered/data/dossiers/DISS-9d8a0574-c000-6534-e044- 00144f67d249/DISS-9d8a0574-c000-6534-e044-00144f67d249_DISS-9d8a0574-c000-6534-e044- 00144f67d249.html. 2-18-2015. Date Accessed 5-26-2015.

3. National Center for Biotechnology Information. PubChem Compound Database; CID=11089; trimellitic anhydride (CAS No. 552-30-7). https://pubchem.ncbi.nlm.nih.gov/compound/11089. 5-16-2015. Date Accessed 5- 21-2012.

4. National Institute for Occupational Safety and Health (NIOSH). Trimellitic anhydride. http://www.cdc.gov/niosh/npg/npgd0635.html. 2-13-2015. Date Accessed 5-22-2015.

5. Organisation for Economic Co-operation and Development (OECD). Trimellitic Anhydride and Trimellitic Acid (CAS Nos 552-30-7; 528-44-9). http://www.chem.unep.ch/irptc/sids/OECDSIDS/TLANA.pdf. 2002. Date Accessed 5-20-0015.

6. Chai OH, Lee H-K, Lee YC, Lee MS, Han E-H, Kim HT, and Song CH. Roles of TNF-α and IgE in the late phase of contact hypersensitivity induced by trimellitic anhydride. Experimental and Molecular Medicine. 2005;37:(5):408-417.

7. Liu FT, Bargatze RF, and Katz DH. Induction of immunologic tolerance to the trimellitate haptenic group in mice: model for a therapeutic approach to trimellitic anhydride-induced hypersensitivity syndromes in humans? J Allergy Clin Immunol. 1980;66:(4):322-326.

8. Fiume MM, Heldreth BA, Bergfeld WF, Belsito DV, Hill RA, Klaassen CD, Liebler DC, Marks Jr JG, Shank RC, Slaga TJ, Snyder PW, and Andersen FA. Final report of the Cosmetic Ingredient Review Expert Panel on the safety assessment of dicarboxylic acids, salts,and esters. Int J Toxicol. 2012;3:(Suppl 1):5S-76S.

9. European Chemical Agency (ECHA). Cyclohex-1,4-ylenedimethanol (cyclohexanedimethanol; CAS No. 105-08- 8). http://apps.echa.europa.eu/registered/data/dossiers/DISS-937e9b7b-a9c8-1736-e044- 00144f67d249/DISS-937e9b7b-a9c8-1736-e044-00144f67d249_DISS-937e9b7b-a9c8-1736-e044- 00144f67d249.html. 4-13-2015. Date Accessed 5-26-2015.

10. National Center for Biotechnology Information. PubChem Compund Database; CID=7735; 1,4- cyclohexanedimethanol (CAS No. 105-08-8). https://pubchem.ncbi.nlm.nih.gov/compound/7735. 5-16- 2015. Date Accessed 5-21-2015.

11. Andersen FA (ed). Special Report: Reproductive and developmental toxicity of ethylene glycol and its ethers. Int J Toxicol. 1999;18:(Suppl 2):53-67.

12. European Chemical Agency (ECHA). Isophthalic acid (CAS No. 121-91- 5). http://apps.echa.europa.eu/registered/data/dossiers/DISS-9d88d245-6ac3-6d31-e044- 00144f67d249/DISS-9d88d245-6ac3-6d31-e044-00144f67d249_DISS-9d88d245-6ac3-6d31-e044- 00144f67d249.html. 4-13-2015. Date Accessed 5-26-2015.

13. National Center for Biotechnology Information. PubChem Compound Database; CID=8496; Isophthalic acid (CAS No. 121-91-5). https://pubchem.ncbi.nlm.nih.gov/compound/8496. 5-16-2015. Date Accessed 5-21-2015.

14. Organisation for Economic Co-operation and Development (OECD). Isophthalic Acid (CAS No. 121-91- 5). http://webnet.oecd.org/HPV/UI/handler.axd?id=a2ab2415-8bdd-4338-88a1-156ba0a24145. 2002. Date Accessed 5-21-2015. Distributed for comment only -- do not cite or quote

15. Environmental Protection Agency (EPA). Maleic anhydride (CAS No, 108-31-6). 10-18-2013. Date Accessed 5-18- 2015.

16. Office of Environmental Health Hazard Assessment (OEHHA). Chronic toxicity summary: maleic anhydride. http://oehha.ca.gov/air/chronic_rels/pdf/maleic.pdf. 2001.

17. Organisation for Economic Co-operation and Development (OECD). SIDS Initial Assessment Profile: Maleic Anhydride. http://webnet.oecd.org/Hpv/UI/handler.axd?id=ab4b7988-908a-4763-939f-e693a5e9c9a7. 2004. Date Accessed 5-18-2015.

18. Organisation for Economic Co-operation and Development (OECD). SIDS Initial Assessment Report for SIAM 18: Maleic Anhydride and Maleic Acid. http://webnet.oecd.org/HPV/UI/SIDS_Details.aspx?key=a7ea6052- cf15-4dca-8866-bd6f1eca33bb&idx=0. 2004. Date Accessed 5-14-2015.

19. Andersen FA (ed). Final report on the safety assessment of maleic acid. Int J Toxicol. 2007;26:(Suppl 2):125-130.

20. European Chemical Agency (ECHA). 2,2-Dimethylpropane-1,3-diol (neopentyl glycol; CAS No. 166-30- 7). http://apps.echa.europa.eu/registered/data/dossiers/DISS-97d93e57-2daa-6bc7-e044- 00144f67d031/DISS-97d93e57-2daa-6bc7-e044-00144f67d031_DISS-97d93e57-2daa-6bc7-e044- 00144f67d031.html. 10-28-2014. Date Accessed 5-27-2015.

21. Organisation for Economic Co-operation and Development (OECD). Neopentyl Glycol (CAS No. 126-30- 7). http://webnet.oecd.org/HPV/UI/handler.axd?id=21fe8103-2fb5-4d44-bc41-5d4fc90a031c. 2002. Date Accessed 5-14-2015.

22. European Chemical Agency (ECHA). Phthalic anhydride (CAS No. 85-44- 9). http://apps.echa.europa.eu/registered/data/dossiers/DISS-9d9f0779-ad0a-28df-e044- 00144f67d249/DISS-9d9f0779-ad0a-28df-e044-00144f67d249_DISS-9d9f0779-ad0a-28df-e044- 00144f67d249.html. 10-28-2014. Date Accessed 5-27-2015.

23. National Center for Biotechnology Information. PubChem Compound Database; CIR=6811; phthalic anhdydride (CAS No 85-44-9). https://pubchem.ncbi.nlm.nih.gov/compound/6811. 5-16-2015. Date Accessed 5-21- 2015.

24. Organisation for Economic Co-operation and Development (OECD). Phthalic anhydride (CAS No. 85-44- 9). http://webnet.oecd.org/HPV/UI/handler.axd?id=ce1be9d2-c97e-414a-b21f-60e9f4923a38. 2006. Date Accessed 5-21-2015.

25. World Health Organization (WHO), Kim JH, Gibb HJ, and Iannucci A. Effects on laboratory mammals and in vitro test systems. Chapter: 8. In: International Programme on Chemical Safety (IPCS). Concise International Chemical Assessment Document 75. Cyclic acid anhydrides: Human health aspects. Geneva, Switzerland: World Health Organization; 2009:

26. Fiume MM, Bergfeld WF, Belsito DV, Hill RA, Klaassen CD, Liebler DC, Marks Jr JG, Shank RC, Slaga TJ, Snyder PW, and Andersen FA. Safety assessment of propylene glycol, tripropylene glycol, and PPGs as used in cosmetics. Int J Toxicol. 2012;31:(Suppl 2):245S-260S.

27. Andersen FA (ed). Final report on the safety assessment of propylene glycol and polypropylene glycols. J Am Coll Toxicol. 1994;13:(6):437-491.

28. European Chemical Agency (ECHA). 2,2,4-Trimethylpentane-1,3-diol (trimethylpentanediol; CAS No. 144-19- 4). http://apps.echa.europa.eu/registered/data/dossiers/DISS-d0199a09-c4c5-4516-e044- 00144f67d249/DISS-d0199a09-c4c5-4516-e044-00144f67d249_DISS-d0199a09-c4c5-4516-e044- 00144f67d249.html. 3-11-2014. Date Accessed 5-26-2015.

29. National Center for Biotechnology Information. PubChem Compound Database; CID=8946: 2,2,4-trimethyl-1,3- pentanediol (trimethylpentanediol; CAS No. 144-19- 4). https://pubchem.ncbi.nlm.nih.gov/compound/2_2_4-trimethyl-1_3-pentanediol. 5-23-2015. Date Accessed 5-26-2015. Distributed for comment only -- do not cite or quote

30. European Chemical Agency (ECHA). Propylidynetrimethanol (trimethylolpropane; CAS No. 77-99- 6). http://apps.echa.europa.eu/registered/data/dossiers/DISS-97d9f45a-4ff2-1b86-e044- 00144f67d031/DISS-97d9f45a-4ff2-1b86-e044-00144f67d031_DISS-97d9f45a-4ff2-1b86-e044- 00144f67d031.html. 4-13-2015. Date Accessed 5-26-2015.

31. National Center for Biotechnology Information. PubChem Compound Database; CID=6510: trimethylolpropane (CAS No. 77-99-6). https://pubchem.ncbi.nlm.nih.gov/compound/Trimethylolpropane. 5-23-2015. Date Accessed 5-26-2015.

32. SAAPedia. Adipic acid/CHDM/MA/neopentyl glycol/trimellitic anhydride copolymer. http://www.saapedia.org/en/saa/?type=detail&id=939. 8-22-2014. Date Accessed 5-14-2015.

33. Lanxess. Technical information sheet: Uniplex 670P (Adipic Acid/Neopentyl Glycol/Trimellitic Anhydride Copolymer). https://techcenter.lanxess.com/fcc/emea/en/products/datasheet/24543303.pdf?docId=2454330 4&gid=24527585&pid=1095. 7-24-2012. Date Accessed 5-14-2015.

34. National Center for Biotechnology Information. PubChem Compound Database; CID 168798: adipic acid/neopenyl glycol trimellitic anhydride copolymer (CAS No. 28407-73- 0). http://pubchem.ncbi.nlm.nih.gov/compound/168798. 2005. Date Accessed 2-25-2014.

35. Huang H, Zhang J, and Shi W. Photopolymerizable polyfunctional methacrylated polyesters. J Photopolym Sci Technol. 1997;10:(2):341-346.

36. Nassif AS, Le Coz CJ, and Collet E. A rare nail polish allergen: phthalic anhydride, trimellitic anhydride and glycols copolymer. Contact Dermatitis. 2007;56:(3):172-173.

37. Food and Drug Administration (FDA). Frequency of use of cosmetic ingredients. FDA Database. 2015.

38. Personal Care Products Council. 4-10-2015. Concentration of Use by FDA Product Category: Trimellitic Anhydride Copolymers. Unpublished data submitted by Personal Care Products Council.

39. European Commission. CosIng database; following Cosmetic Regulation No. 1223/2009. http://ec.europa.eu/consumers/cosmetics/cosing/. 2015. Date Accessed 2-27-2015.

40. Gach JE, Stone NM, and Finch TM. A series of four cases of allergic contact dermatitis to phthalic anhydride/trimellitic anhydride/glycols copolymer in nail varnish. Contact Dermatitis. 2005;53:(1):63-64.

41. Moffitt DL and Sansom JE. Allergic contact dermatitis from phthalic anhydride/trimellitic anhydride/glycols copolymer in nail varnish. Contact Dermatitis. 2002;46:(4):236.

Distributed for comment only -- do not cite or quote

Special Report: Reproductive and Developmental Toxicity of Ethylene Glycol and Its Ethers1

Hydrogenated Lanolin; PEG-75 Lanolin Oil; and PEG-75 Polymers of Ethylene Glycol are linked via ether linkages with Lanolin Wax (CIR 1996f). various alcohols or via ester linkages to various fatty acids in many PEG-5, -10,-16,-25, and -40 Soya Sterol (CIR 1996g). cosmetic ingredients. Ethylene Glycol, when reacted with an alkyl alcohol, forms an ethylene glycol monoalkyl ether. These com To provide a perspective on the chemical structures of the pounds are metabolized in the human body by alcohol dehydro compounds involved, and to form a basis for comparison with genase and aldehyde dehydrogenase to form corresponding ac the structures of the PEG-derivatives listed above, this special etaldehyde and acetic acid derivatives. Data are presented that show reproductive and developmental toxicity is associated with report begins with a presentation of the structure of ethylene metabolites of ethylene glycol monoalkyl ethers, but not with the glycol and its monoalkyl ethers. monoalkyl ethers themselves. Further, it is suggested that the tox Ethylene Glycol, when reacted with an alkyl alcohol, a corre icity of these metabolites is inversely proportional to the length of sponding ethylene glycol monoalkyl etheris formed, as shown in the alkyl chain in the original alkyl ether. In the case of the com Figure 1. Ethylene Glycol Monomethyl Ether and Ethylene Gly pounds used in cosmetics, most have alcohols or fatty acids linked to polyethylene glycol chains, not a single Ethylene Glycol moiety. col Monoethyl Ether are, respectively, 2-Methoxyethanol and Where Ethylene Glycol is linked to a fatty acid by an ester linkage, 2-Ethoxyethanol. The ethylene glycol monoalkyl ethers are me the resulting compound is chemically different from the monoalkyl tabolized by alcohol dehydrogenase and aldehyde dehydroge ethers. Where Ethylene Glycol is linked to an alcohol via an ether nase, resulting in aldehyde and acetic acid derivatives as shown linkage, the alkyl chain is large and complex, suggesting little or in Figure 2. Ifthe initial compound is 2-methoxyethanol, then the no potential toxicity. Overall, it was found that metabolites of ethy lene glycol monoalkyl ethers are reproductive and developmental metabolites are methoxyacetaldehyde and methoxyacetic acid. toxins. In general, however, the metabolites of concern are not ex This special report includes data on Ethylene Glycol, 2- pected to be formed in cosmetic formulations that contain polymers Methoxyethanol, and 2-Ethoxyethanol (the latter two are under of ethylene glycol. the section Ethylene Glycol Ethers). 2-Butoxyethanol, a less harmful reproductive and developmental toxin, has been pre viously reviewed by the CIR Expert Panel which found that, In their review of the safety assessments of various polyethy in oral and inhalation reproduction toxicity studies in rodents, lene glycol (PEG) derivatives, the Cosmetic Ingredient Review decreases in the number of viable litters, increases in resorp (CIR) Expert Panel expressed concern about the teratogenicity tions, and ventricular septal defects and arterial defects have and testicular toxicity of the monomer, Ethylene Glycol, and its been reported. Dermal application failed to elicit any reproduc monoalkyl ethers. Safety assessments of PEG-derived cosmetic tive toxicity (Andersen 1996). ingredients for which this concern is an issue include:

Ceteth-1, -2, -3, -4, -5, -6,-10,-12, -14, -15,-16,-20,-24, -25, EXPOSURE LIMITS -30, and -45 (CIR 1996a). The Occupational Safety and Health Administration (OSHA) Oleth-2, -3, -4, -5, -6, -7, -8, -9,-10,-12,-15,-16,-20,-23,-25, and the American Conference of Governmental Industrial Hy -30, -40, -44, and -50 (CIR 1996b). gienists (ACGIH) have established exposure limits for Ethy PEG-2, -3, -5,-10,-15, and -20 Cocamine (CIR 1996c). lene Glycol, 2-Methoxyethanol, and 2-Ethoxyethanol (National PEG-2, -3, -4, -8, -9, -12, -20, -32, -75, -120, -150, and -175 Institute for Occupational Safety and Health [NIOSH] 1983; Distearate (CIR 1996d). ACGIH 1986). The OSHA standards were based primarily on PEG-7, -30,-40,-78, and -80 Glyceryl Cocoate (CIR 1996e). reports of blood, hepatic, renal, and central nervous system tox PEG-5, -10,-20,-24,-25,-27,-30,-35,-40,-50,-55,-60,-75, icity caused by 2-Methoxyethanol and 2-Ethoxyethanol; no re -85, and -100 Lanolin; PEG-5, -10, -20, -24, -30, and -70 productive studies were considered when the standards were adopted. The ACGIH threshold limit values for 2-Methoxyetha nol and 2-Ethoxyethanol were changed to the present 5 ppm Received 25 February 1999; accepted 12 May 1999. limits from 25 and 50 ppm, respectively, based on evidence 1Reviewed by tbe Cosmetic Ingredient Review Expert Panel. Re becca S. Lanigan, former Scientific Analyst and Writer, prepared this of testicular toxicity. The ACGIH threshold limit value for 2- report. Address correspondence to Dr. F. Alan Andersen, Director, CIR, Butoxyethanol is included for the sake of comparison. These 110117tb Street, NW, Suite 310, Washington, DC 20036, USA. limits are given in Table 1 (NIOSH 1983).

54 COSMETIC INGREDIENT REVIEW

OH-CHz-CHz-OH ethylene glycol or 3333 mg/kg was adminstered subcutaneously, either alone or simulaneously with an oral dose of 530 mg/kg aqueous NaHC03 + (an endogenous agent known to correct metabolic acidosis). In all studies, drinking water was supplemented with 2.65 mg/ml R-OH alkyl alcohol NaHC03. Blood samples hadhyperosmolality, metabolic acido sis, and dose-related increases in the anion gaps (comparable to controls) and osmolar gap (returned to normal at 9 hours). Uri nalysis results included dose-related diuresis and a significant HO-C~-C~-0-R ethylene glycol monoalkyl ether decrease in osmolality without significant changes in pH, Na+, and/or K+ values. Urine sediment from all groups (including FIGURE 1 controls) contained calcium oxalate crystals, which were more Reaction of Ethylene Glycol with an alkyl alcohol to form an prevalent in the high dose groups. ethylene glycol monoalkyl ether. In the second study, a teratology study, 2800 or 3333 mg/kg Ethylene Glycol doses were injected subcutaneously with or NIOSH proposed that 2-Methoxyethanol and 2-Ethoxyetha without NaHC03 on GD 11 using 10--15 dams per treatment nol have the potential to cause adverse reproductive effects group. Developmental effects were evaluated in term fetuses in male and female workers. This recommendation was based (GD 22); two thirds were examined for skeletal defects and the on published studies in which dose-related embryotoxicity and remainder for visceral anomalies. Three of 13 dams treated with other reproductive effects were found in a number of animal 3333 mglkg Ethylene Glycol alone died, but 2800 mg/kg was not species exposed to the chemicals via different routes of admin maternolethal. In the 2800 mglkg dose group, 55/136 fetuses istration; of particular concern were studies in which adverse had skeletal anomalies after Ethylene Glycol treatment alone. effects were observed in pregnant animals exposed to concen With NaHC03 administration, this number was 20/128. At the trations below the government standards. As a result, NIOSH higher dose, the values for skeletal anomalies were 70/82 and suggested that worker exposure be minimized whenever possi 46/83, respectively. Control values were 111106 (NaHC03) and ble (NIOSH 1983). 4/110 (water). Simultaneous treatment with NaHC03 signifi cantly reduced the maternal toxicity, loss in fetal body weight, ETHYLENE GLYCOL and the incidence of fetal skeletal defects. Malformations in Khera (1991) investigated the possible relationships between cluded retarded ossifications in sternebrae, vertebrae, metacar changes in maternal homeostasis and the incidence of fetal anom pals, and metatarsals, and a reduced number of ribs (eight to 12) alies in pregnant Sprague-Dawley rats induced by maternotoxic or fused ribs. doses of Ethylene Glycol. The first of three studies was an acid In the third study, conceptuses in situ, still enclosed in the base electrolyte study on gestational day (GD) 11 using five uterine capsule, were examined. A single dose of 3333 mg/kg to eight cannulated rats per treatment group. Doses of 1250, Ethylene Glycol alone or with NaHC03 was given on GD 11 2500, or 5000 mg/kg Ethylene Glycol were administered orally or multiple doses of 5000 mglkg/day Ethylene Glycol were ad ministered orally from GD 7-13. Dams (6 to 12 per group) HO-c~-c~-o-R ethylene glycol monoalkyl ether were killed 24 or 48 hours after the single or last dose of Ethy lene Glycol. Karyorrhexis and pyknosis of mesodermal cells in the allantoic bulb, an absence of chorioallantoic fusion, pla cental thrombosis, and maternal hemorrhage in the yolk sac alcohol dehydrogenase cavity were detected (3333 mglkg Ethylene Glycol alone). No lesions were found in the conceptuses and dams of control

rats (water or NaHC08 alone). The total cross-sectional area o=cH-CHz-0-R aldehyde derivative of maternal vascular spaces increased following treatment with Ethylene Glycol, whereas that of the fetuses decreased. Simul taneous treatment with NaHC03 increased the surface area of aldehyde dehydrogenase gaseous exchange between the apposing maternal and fetal vas cular channels. The higher dose of 5000 mg/kg Ethylene Gly col caused lesions in all but one conceptus, and reductions in the width of the labyrinth (consisting of larger maternal spaces o=C--CH:!-0-R acetic acid derivative I and proportionately smaller and fewer allantoic villi) and the OH basal zone of the placenta were observed. Additionally, two of four conceptuses had intraplacental hematomas from a maternal FIGURE 2 hemorrhage. One such hematoma contained crystals morpho Metabolism of ethylene glycol monoethers. logically similar to those of calcium oxalate. Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 55

TABLEt Exposure limit values (NIOSH 1983)

ACGIH threshold OSHA permissible Compound limit valuea exposure limita

Ethylene Glycol (vapor) 50 ppm (125 mg!m3) 2-Methoxyethanol 25 ppm (80 mg/m3)b 2-Ethoxyethanol 5 ppm (19 mg/m3)b 200 ppm (740 mg/m3)b 2-Butoxyethanol 25 ppm (120 mg/m3l

a values as time-weighted averages for 8-hour work shifts during a standard 40-hour week. h"Skin" notation, indicating the potential for skin absorption of toxic amounts.

The maternal metabolic acidosis and hyperosmolality de formance. At 70 days of age, 20 F1 offspring per sex from the tected in the first study were suspected to contribute towards two groups were randomly paired and mated with nonsiblings, the reductions in villigenesis, chorion size, and erythroblastic and the resulting litters were necropsied. population in the allantois, as well as the increase in the in No clinical signs of Ethylene Glycol toxicity were observed cidence of fetal abnormalities and reduced fetal body weight in the F0 generation; however, two females given 0.5% Ethylene observed in the second and third studies. Materno-embryonic Glycol and one male and two females in the control group died gaseous exchange and fetal nutrition were reduced, impairing prematurely. At least one death in the 0.5% Ethylene Glycol fetal development, by maternal factors due to treatment with group might have been due to the treatment, as the renal tubules Ethylene Glycol, including metabolic acidosis, hyperosmolality, contained a moderate number of oxalate crystals. Exposure to hemorrhages in the ectoplacental cone, extraembryonic cavities, 1% Ethylene Glycol was associated with significant decreases and around Reichert's membrane, and necrosis of the decidua in the number of litters (F1) per fertile pair, the mean number of basalis (Khera 1991). live births per litter, and the mean live pup weight as compared Lamb et al. (1985) administered Ethylene Glycol in drink to controls (Table 2). ing water to male and female CD-1 albino mice in a continuous In the assessment of reproductive performance for the F1 gen breeding program to determine reproductive function and off eration, fertility was 80% and 61% for control and 1% Ethylene spring development. Fo mice were treated with 0.25, 0.5, or 1.0% Glycol treated mice, respectively; the difference was not statis (wlv) Ethylene Glycol. Forty and 20 mice per sex made up the ve tically different, however. The lower live birth index and live hicle control group and each treatment group, respectively. Mice pup weight were also not significant. No clinical signs of Ethy were treated during a 1-week premating period after which the lene Glycol toxicity were observed, but some offspring (F2) had male and female mice were randomly paired within each dose shorter snouts and wide-set eyes compared to controls, as well as group and cohabited for 14 weeks while treatment continued. shortened frontal, nasal, and parietal bones, one or more pairs of Newborn litters were evaluated and immediately killed by de fused ribs, abnormally shaped or missing sternebrae and/or ver capitation. At the end of the 14 weeks, males and females were tebrae, and twisting of the spine. Bones were smaller and had separated and the litters saved during the following 3-week pe altered shapes compared to controls under low-magnification riod. Treatment was also continued during this time and until examination; however, no lesions were found at microscopic the end of the study. Final litters (F 1) from control and high examination. Calcium oxalate crystals were not observed in the dose groups were weaned and evaluated for reproductive per- renal tubules of either group. The researchers concluded that

TABLE2 Reproductive performance of mice given Ethylene Glycol (Lamb et al. 1985)

Treatment Pairs with Surviving Litters Live pups Proportion of pups Live pup group litters pairs per pair per litter born alive weight (g) Control 40/40 38/40 4.9 ± 0.08 10.8 ± 0.5 0.96 ±0.03 1.63 ± 0.02 0.25% 20/20 20/20 4.7 ± 0.02 10.4 ± 0.5 0.96 ±0.02 1.64 ± 0.02 0.5% 20/20 18/20 4.9 ± 0.08 10.5 ±0.6 0.97 ± O.Ql 1.58 ± 0.02 1.0% 20/20 20/20 4.5 ±0.2b 10.2 ± 0.3a 0.96±0.02 1.53 ± 0.02b

a significantly different from control (p < .05). hSignificantly different from control (p < .01). Distributed for comment only -- do not cite or quote

56 COSMETIC INGREDIENT REVIEW

TABLE3 Teratologic evaluation of Ethylene Glycol-CD-1 mice (NCTR 1984a)

Ethylene %Adversely % Litters with % live litters with Glycol dose affected ::::1 affected % Malformed live ::::1 malformed (mglkg/day) implants/litter implant fetuses/live litter live fetus Control 11.17 72.0 0.25 4.0 750 19.65 83.3 10.0 66.67 1500 50.67 95.7 37.77 81.82 3000 64.41 100 56.54 95.65

Ethylene Glycol was a "weak reproductive toxicant, but a po (10 mllkg) were given in water on GD 6-15 to 27-29 CD rats tential teratogen" (Lamb et al. 1985). per group. No unscheduled maternal deaths or distinctive clinical In another study using CD-1 mice, researchers administered signs were observed other than piloerection. Dose-dependent de 750, 1500, or 3000 mg/kg/day Ethylene Glycol in drinking water creases in maternal body weight, body weight gain, liver weight, (10 ml/kg/day dose volume) to 23-25 pregnant mice per group and relative kidney weights were observed. These changes were on GD 6-15. The mice were killed on GD 17. No unscheduled significant in the mid- and high-dose groups. No significant maternal deaths or toxicity signs were observed except for pi differences were found between treatment groups and vehicle loerection and dose-related decreased body weight gain, body control groups in the number of corpora lutealdam, number of weight loss on GD 11-17, and decreased liver weight. The body implantation sites/dam, percentage of preimplantation loss, or weight loss was significant in the mid- and high-dose groups. At the percentage of dead fetuses. However, the percent of nonlive a dose of 1500 mg/kg/day Ethylene Glycol, the dams had a lower implants/litter was increased in the high-dose group, and 21.0% average number of corpora lutea and a decrease in the number of resorptions per litter occurred (compared to 4.7% for controls; implantation sites/litter. The high dose caused 19.86% resorp see Table 4). Common malformations included clefts of the lip, tions/litter (compared to 8.74% in the control group). A dose face, and palate, and varying degrees of skeletal dysplasia (de related increase occurred in the number of resorptions per litter, fects of the cranium, ribs, and vertebral column) (NCTR 1984b ). and a significant decrease in the incidence of dead fetuses (dose A study conducted by the National Toxicology Program dependent) at the top two doses (Table 3). In general, significant (NTP) reported that the no observable adverse effect levels increases were found at the mid- and high-doses in the per (NOAEL) for maternal and developmental toxicity were 1000 centage of adverse effects in conceptuses/litter during the post and 2000 mg/kg/day Ethylene Glycol, respectively, when ad implantation phase of pregnancy. Fetal body weight dropped in a ministered to artificially inseminated New Zealand white rab dose-dependent manner. The most common fetal abnormalities bits (NTP 1991). Doses of 100, 500, 1000, or 2000 mglkg/day observed were hydronephrosis, craniofacial anomalies (clefts Ethylene Glycol (in water) were given by gavage on GD 6-19. of the face, lip, and palate) and skeletal dysmorphogenesis that The dose volume was 5 ml/kg body weight. Live fetuses were included fused, short, branched or missing ribs, fused or mis dissected from the uterus on GD 30 after examination of the ma aligned sternebrae, incomplete ossification, and fused thoracic ternalliver, kidneys, intact uterus, and counting of the corpora or lumbar arches (National Center for Toxicological Research lutea. At the 2000 mglkg dose, 42.1% (8/17) of does were killed [NCTR] 1984a). on GD 9-25 due to acute renal failure, the pregnancy rate was In a second study by NCTR (1984b), the results were simi 81.8%, three does delivered early, and one doe aborted on GD 20. lar. Doses of 1250, 2500, or 5000 mglkg/day Ethylene Glycol The only significant effect of treatment observed at necropsy was

TABLE 4 Teratologic evaluation of Ethylene Glycol-CD rats (NCTR 1984b)

Ethylene %Adversely % Litters with % live litters Glycol dose affected ::::1 affected % Malformed live with:::: 1 (mglkg/day) implants/litter implant fetuses/live litter malformed live fetus Control 6.02 42.9 1.37 7.14 1250 12.02 67.9 6.65 39.29 2500 29.52 89.7 25.11 68.97 5000 76.59 96.3 73.53 96.15 Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 57 a slight increase in maternal absolute kidney weight. Upon fur sensitivity was rabbits > mice > rats, but for developmental ther examination, renal lesions of the cortical tubules included toxicity the order was mice » rats »rabbits (NTP 1991; Tyl intraluminal crystals, and epithelial necrosis. Pregnancy rates et al. 1993). were 95.5, 95.7, 91.3, and 95.2% for control and three treatment Pregnant Fischer 344 rats were treated with 0.2, 0.4, and groups (up to 1000 mg/kg/day), respectively. One doe delivered 1.0 g/kg/day Ethylene Glycol in the diet on GD 6-15 and killed early in each dose group. Except for a significant increase in on GD 21. A positive-control group received 500 mg/kg hydrox the number of implantation sites/litter and a slight increase in yurea in saline intraperitoneally on GD 11. The negative-control the live litter size at 500 mg/kg/day, no treatment-related effects group received stock diet. Fetuses were delivered by caesarian on any gestational parameters were observed. However, the val section on GD 21 and examined for teratogenic effects. Clinical ues at that dose were within historical control values. Treatment signs of maternal toxicity were not observed, and no significant with Ethylene Glycol did not result in significant fetotoxicity or differences were detected between control and treated mater increase the incidence of fetal malformations at any dose tested nal weight gain on GD 6-21. Fetal length, weight, the number using New Zealand white rabbits. of total implantations, and litter size were not different between The same researchers compared various toxicity endpoints to controls and treated animals. Preimplantation loss was greater at determine the relative sensitivities of New Zealand white rabbits, 1.0 g/kg/day Ethylene Glycol, but the difference was not signifi Sprague-Dawley rats, and Swiss mice to oral doses of Ethylene cant. No increased incidences of major fetal anomalies, skeletal Glycol during organogenesis (Table 5). For maternal toxicity, or visceral, were observed after treatment with Ethylene Glycol;

TABLES NOAELS for Ethylene Glycol

Route and time of Species administration NOAEL (mg/kg/day) Reference Maternal Toxicity New Zealand white rabbit Gavage on GD 6-19 1000 Tyl et al. 1993 Sprague-Dawley rat Gavage on GD 6-15 < 1250 (not determined) Price et al. 1985 Fischer 344 rat Dosed feed on GD 6-15 1000 Maronpot et al. 1983 CD rat Inhalation on GD 6-15 1000 Tyl et al. 1995b (whole-body) CD rat Gavage on GD 6-15 1000 Neeper-Bradley et al. 1995 CD-1 mouse Gavage on GD 6-15 750 Price et al. 1985 CD-1 mouse Drinking water 400 Lamb et al. 1985 (multigenerational) CD-1 mouse Dermal application on >3549 Tyl et al. 1995a GD6-15 CD-1 mouse Inhalation on GD 6-15 150 mg/m3 (whole-body); Tyl et al. 1995b Swiss mouse Gavage on GD 6-15 1500 Tyl et al. 1989 Developmental Toxicity New Zealand white rabbit Gavage on GD 6-19 2000 Tyl et al. 1993 Sprague-Dawley rat Gavage on GD 6-15 <1250 (not determined) Price et al. 1985 CD rat Gavage on GD 6-15 500 Neeper-Bradley et al. 1995 CD rat Inhalation on GD 6-15 150 mg/m3 (whole-body) Tyl et al. 1995b Fischer 344 rat Dosed feed on GD 6-15 200 Maronpot et al. 1983 CD-1 mouse Gavage on GD 6-15 150 Neeper-Bradley et al. 1995 CD-1 mouse Gavage on GD 6-15 <750 (not determined) Price et al. 1985 CD-1 mouse Drinking water 800 Lamb et al. 1985 (multigenerational) CD-1 mouse Dermal application on 3549 Tyl et al. 1995a GD6-15 CD-1 mouse Inhalation on GD 6-15 < 150 mg/m3 (whole-body) Tyl et al. 1995b Swiss mouse Gavage 150 Tyl et al. 1989

Note. GD, gestational day. Distributed for comment only -- do not cite or quote

58 COSMETIC INGREDIENT REVIEW however, incidences of poorly ossified and unossified vertebral occurred after treatment with 1000 mg/m3 and 2500 mg/m3. centra significantly increased in fetuses of the high dose group Signs of embryo/fetal toxicity included an increase in nonviable (Maronpot et al. 1983). implantations/litter, a reduction in viable implantations/litter, Ethylene Glycol (in water) was applied to the skin of pregnant reduced fetal body weights/litter, increases of individual and CD-1 mice (30 per group) on GD 6-15 at dosages of 404, 1677, pooled malformations (external, visceral, and skeletal), and an and 3549 mg!k:g/day. The positive control was 3000 mg/kg/day increase of total malformations. Observed malformations in Ethylene Glycol. Ethylene Glycol was applied to the clipped cluded exencephaly, cleft palates, foreshortened and abnormal and shaved skin of the dorsum, below the scapulae caudal to the faces or facial bones, vertebral fusions, and forked, fused, or neck. Each test solution was applied as a 100-pJ volume. The missing ribs. The maternal NOAEL was 150 mg/m3 and the application site was then covered with sterile gauze and masking developmental NOAEL was 150 mg/m3 (Tyl et al. 1995b). tape. Mice of the positive control group received 3000 mg/kg/day The mechanism of Ethylene Glycol toxicity is not fully un (300 mg/ml; 10 mllkg) Ethylene Glycol by gavage. Dams were derstood. In all cases, however, toxicity of Ethylene Glycol ap killed for necropsy on GD 18. Fetuses were removed, weighed, peared to be due to its metabolism to more toxic compounds such sexed, and examined for signs of developmental toxicity. Dams as glycoaldehyde and glycolic acid by hepatic alcohol dehydro given dermal applications of Ethylene Glycol had no treatment genase (primarily) and other enzymes (Jacobsen and McMartin related maternal toxicity, differences in pre- or postimplantation 1986; NTP 1993a). When enzyme activity was blocked, toxicity loss, fetal body weights/litter, or increased incidences of fetal did not occur (Jacobsen and McMartin 1986; Khera 1991). malformations. Dams of the positive control group had increased Carney et al. (1996) identified glycolic acid as the proxi water consumption, reduced fetal body weights/litter, and in mate toxicant for Ethylene Glycol developmental toxicity us creased fetal malformations and variations. Eight dams given ing CD rat whole embryo cultures. In this study, the 10.5-day Ethylene Glycol by gavage died before scheduled necropsy. The conceptuses (10 per group) were cultured and exposed to 0.5- kidneys of mice of the positive control group had tubular cell de 50.0 mMIL of Ethylene Glycol or glycolic acid for 46 hours. generation, but no oxalate crystals. Minimal renal tubular lesions Other embryos were cultured in media containing 12.5-50.0 mM were observed in three mice. The maternal and developmental glycolic acid or sodium glycolate. The positive control was NOAELS were both 3549 mg/kg/day (Tyl et al. 1995a). 1.0 mM sodium valproate. The lowest effect level of Ethylene Tyl et al. (1995b) administered a respirable aerosol of 150- Glycol was 50.0 mM/L, which caused a very slight decrease in 2500 mg/m3 Ethylene Glycol (whole-body exposure) to preg morphology score. Glycolic acid caused a significant increase in nant CD rats and CD-1 mice on GD 6-15 (6 hours per day). the percentage of dysmorphic embryos. Disorganized patterns of Rats were killed on GD 21 and mice on GD 18. Dams were visceral yolk sac vessels were also observed. When glycolic acid evaluated for body weight, the weights of the gravid uterus, and sodium glycolate were added to the medium, the percentage kidneys, and liver, the number of ovarian corpora lutea, and sta of embryos with active yolk sac circulation was slightly, but not tus of implantation sites (i.e., number of resorptions, dead and significantly decreased. Embryo viability was not affected. A live fetuses). Fetuses were dissected from the uterus, counted, significant increase in the percentage of dysmorphic embryos weighed, sexed, and examined for external, visceral, and skele was observed after treatment with 12.5 mM glycolic acid and tal malformations. All rat dams survived to scheduled necropsy. sodium glycolate. The high dose caused a significant increase in absolute and rel The two major routes of potential human exposure to Ethy ative liver weight, but other parameters were normal. No other lene Glycol are dermal and inhalation routes. However, inves signs of maternal toxicity were observed. Gestational parame tigators determined that Ethylene Glycol was poorly absorbed ters (i.e., pre- and post-implantation loss, live fetuses/litter, sex through the skin and its low vapor pressure made significant in ratio, and fetal body weight/litter) did not differ from controls. halation exposures less likely. Investigators concluded that nor Treatment-related increases did not occur in the incidence of mal human uses of Ethylene Glycol would result in negligible individual malformations, pooled external, visceral, or skele plasma concentrations of Ethylene Glycol that were well below tal malformations, or total malformations by fetus or litter. No the threshold limits for reproductive and developmental toxicity increases occurred in the incidence of external or visceral vari (Carney 1994; Sun, Frantz, and Beskitt 1995). ations. Reduced ossification of the humerus, zygomatic arch, metatarsals, and proximal phalanges of the hindlimb were ob ETHYLENE GLYCOL ETHERS served in fetuses of rat dams given 1000 mg/m3 and 2500 mg/m3 The monoalkyl ethers of Ethylene Glycol cause reproduc Ethylene Glycol. The maternal NOAEL was 1000 mg/m3, and tive and developmental toxicity (NTP 1993b). Studies have sug the developmental NOAEL was 150 mg/m3 (Tyl et al. 1995a). gested that the metabolites of2-Methoxyethanol and 2-Ethoxye All mouse dams survived to scheduled necropsy. One dam thanol (such as methoxyacetaldehyde and methoxyacetic acid) given the high dose had a totally resorbed litter. Dams given the are the toxic agents and the proximate teratogens (Foster et al. 1000 mg/m3 and 2500 mg/m3 Ethylene Glycol had reduced body 1984; Miller et al. 1984; Gray et al. 1985; Ritter et al. 1985; weight and weight gain during and after the exposure period, as Beattie and Brabec 1986; Miller 1987; Scott et al. 1989; Clarke, well as reduced gravid uterine weight. Embryo/fetal toxicity Duignan, and Welsch 1990; Chiewchanwit and Au 1994). The Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 59 metabolites (but not the parent monoalkyl ethers) attack rapidly Oral Toxicity dividing cell systems and other tissues with high rates of res Oudiz, Walsh, and Wiley (1993) used a mouse chimera as piration and energy metabolism, often inhibiting mitochondrial say to detect a decrease in cell proliferation of preimplanta respiration (Beattie and Brabec 1986). Methoxyacetic acid is tion embryos as a result of treatment with 2-Methoxyethanol. more toxic than ethoxyacetic acid. These metabolites have the Male Swiss ICR mice were administered 750 mglk:g/day 2- structure RO(CH2)nCOOH. Generally, an increase in n tends Methoxyethanol in water by gavage for 5 days, 12 received to decrease the fetotoxicity of the compound more than an in 1,500 mg/k:g/day 2-Methoxyethanol, and 22 control mice re crease in the -R chain length. For example, minor metabo ceived only distilled water. The males were mated to unex lites of 2-Methoxyethanol such as 4-methoxybutyric acid or posed females during weeks 1-7 posttreatment and embryos 3-methoxypropionic acid are much less active than methoxy were retrieved 26 hours after mating. High-dose males had tran acetic acid in terms of embryotoxicity (Miller 1987). sient weight loss that returned to control values by the first Monoalkyl ethers of Ethylene Glycol are biotransformed to week of the mating period. During week 4, mice treated with their acetaldehyde and acetic acid forms by liver alcohol and 750 mg/k:g/day 2-Methoxyethanol had infertility. The same ef aldehyde dehydrogenases (Foster et al. 1984; Gray et al. 1985; fects were noted in the high-dose group during weeks 3-6. The Chiewchanwit and Au 1994). In treated rats during labelling mean proliferation ratios of surviving four-cell embryos deter studies, the 2-Ethoxyethanol ether linkage was cleaved to pro mined during the chimera assay were significantly less than duce 14C02. Of the dose administered, 11.7% (ethoxy-labelled) the control ratio (0.51 ± 0.04) in the 750 mglkg/day group for or 4.6% (ethanol-labelled) was eliminated as expired 14C02 week 4 (0.41 ± 0.07) and the 1,500 mglkg/day group for week (Cheever et al. 1984). Small amounts of Ethylene Glycol have 5 (0.46 ± 0.04). No dose-related differences were seen in the also been detected in the urine of F344/N rats after inhalation total chimera number seen (Oudiz, Walsh, and Wiley 1993). of the glycol ethers, indicating that the monomer is a minor In a second experiment within the same study, 16 mice each metabolite (Kennedy, Chang, and Henderson 1993). made up the control and treatment groups (50 and 200 mglkg/day Dividing spermatocytes commonly are affected by treatment for 5 days). The males were mated as above. Mean body weights with 2-Methoxyethanol and 2-Ethoxyethanol (and their metabo were not significantly altered compared to the control group lites), resulting in severe testicular toxicity (Foster et al. 1983; (distilled water). Infertility also did not occur in the dosed mice. Chapin and Lamb 1984; Chapin et al. 1984; Gray et al. 1985). The mean proliferation ratios were significantly less for both the Creasy et al. (1985) reported that 250 mglk:g 2-Methoxyethanol 50 and 200 mglkg/day groups during week 4 (0.49 ± O.Ql and (single oral dose) did not affect Sprague-Dawley rat zygotene 0.48 ± 0.02, respectively) compared to that of the control group spermatocytes in stage XIV of meiosis, but after one day of ex (0.50 ± 0.01). No dose-related differences in total cell number posure, 70% of pachytene spermatocytes in stage I were killed or of chimera were observed. The week 4 proliferation ratio de depleted. Dividing spermatocytes were similarly affected whe creases in this and the previous experiment corresponded to the reas step I spermatids were not. In tubules containing large pachytene spermatocyte stage. Overall, changes in sperm were numbers of necrotic cells, fine vacuolation of the Sertoli cell transmitted and expressed within the preimplantation embryo as cytoplasm was observed. Otherwise, Sertoli cells were not af a cell proliferation disadvantage in the chimera assay, even at ex fected. A gradual reduction in susceptibility occurred toward posure concentrations below those that adversely affect fertility midpachytene, and cells in stages VII-XI were unaffected by (Oudiz, Walsh, and Wiley 1993). treatment (Creasy et al. 1985). Scott et al. (1989) treated pregnant Macaca fascicularis pri The glycol ethers inhibit intercellular communication. Doses mates daily with 2-Methoxyethanol during organogenesis to as of 17.5, 20, and 25 JLVml2-Methoxyethanol (0.13-0.3 M) ap sess the embryotoxic effects of the ether. Doses of 12, 24, or parently blocked the function of gap junctions in human palatal 36 mglkg/day 2-Methoxyethanol were administered to four pri mesenchyme cells (HEPM) (Welsch and Stedman 1984a). Cell mates once daily from GD 20-45 by gavage in 15 ml water cell communication was also inhibited in Chinese Hamster V79 through a stomach tube. Two control groups were used (ethanol cells by 2-Methoxyethanol and 2-Ethoxyethanol (Welsch and and untreated). The formation and elimination of methoxyacetic Stedman 1984b). Pulse-labeling and autoradiographic analysis acid were followed in maternal blood samples. Hysterotomies results indicated that 2-Methoxyethanol had adverse effects on were performed on GD 100. Fetuses were examined for gross ab the treated HEPM cells. The cells were rounded, stained more normalities, fixed, x-rayed, dissected, and assayed for methoxy intensely, and had a lower cell density than control culture cells. acetic acid in pharmacokinetic studies. At 12 and 24 mglkg, 3 of Inhibition of gap junction-mediated chemical messenger trans 10 and 3 of 13 embryos died as a result of treatment, respectively. fer did not occur as deduced from study of electron micrographs; The dead embryos were retained in the uterus with minimal em rather, morphological effects suggested that gap junctions were bryonic autolysis. One spontaneous abortion occurred in each not established at sites of cell-cell contact. Because intercellular of these two treatment groups. Dose-related maternal anorexia communication is a fundamental event for the differentiation of (slight to severe) and weight loss were recorded as well. Un embryonal tissues, the researchers hypothesized that this block treated control monkeys had a mean weight of 3.45 kg at the start ing action disrupts fetal organogenesis (Welsch and Stedman of treatment that increased to 3.47 kg by the end of treatment. 1984a). The alcohol control group mean weight increased from 3.53 to Distributed for comment only -- do not cite or quote

60 COSMETIC INGREDIENT REVIEW

3.63 kg. The 3.79, 3.48, and 4.25 kg mean weights of the 12, sorbed, and none were malformed. At the low dose, 105 (i.p.) and 24, and 36 mglkg/day treatment groups decreased to 3.72, 3.28, 101 (oral) implants were counted, respectively; 4.7 and 19.3% and 3.88 kg, respectively. The high dose, 36 mglkg/day, caused were dead and resorbed; and 62.8 and 45.2% of survivors were 100% embryo mortality (8/8) between GD 27-63 and one fetus malformed. The high dose had 106 and 105 implants; 100% was missing a digit from each forelimb. Fetuses that survived to of survivors were malformed; and 20.1 and 15.1% were dead GD 100 had no malformations or growth retardation. Methoxy and resorbed. Eighty to 100% of the malformations were hy acetic acid accumulated in maternal serum after repeated daily dronephrosis (decreased with increasing dose) and heart, tail, dosing. Results of transplacental studies indicated that the con and limb defects. Miscellaneous anomalies included microph centration of the metabolite was uniformly distributed in the thalmia, micrognathia, folded retina, cleft lip, and diaphragmatic embryo and extraembryonic fluids at a concentration similar to hernia. The most common cardiac malformation was interven that in maternal serum and accumulated in the yolk sac at high tricular septal defect, with the lesion in the muscular portion of concentration (Scott et al. 1989). the septum. Other heart defects were levocardia, defects between In a study by Fosteretal. (1983), oral doses (in water) of 100- the ascending aorta and right ventricle, truncus communus, di 500 mglkg/day 2-Methoxyethanol and 250-1000 mglkg/day 2- lated ductus arteriosus, and dilated aortic arch. Short or kinked Ethoxyethanol were administered to male Sprague-Dawley rats tails occurred frequently. Limb defects, including short limb, over 11 days. Sixteen hours after a single 500 mglkg dose of bowed radius, and ventral polydactyly were observed and their 2-Methoxyethanol was given, mitochondrial damage was ob incidence increased with the dose (Ritter et al. 1985). served in spermatocytes. The degeneration of pachytene sper In a 2-week study in which male F344/N rats were treated matocytes was seen within 24 hours after a single 100 mglkg 2- with 200, 400, 600, 1000, or 2000 mglkg/day 2-Methoxyethanol Methoxyethanol dose. Similar results occurred at day 11 when in drinking water, the absolute right testis weight was 1.235 for 500-1000 mglkg/day 2-Ethoxyethanol doses were given. Chro control and 1.182, 0.667, 0.429, 0.372, or 0.316 g for dosed matin margination of spermatid nuclei occurred by day 4 groups, respectively. The respective relative testes weights were after a 500 mg/kg/day dose of 2-Methoxyethanol and day 7 6.07 for control and 5.59, 3.29, 2.38, 2.51, and 2.35 g for dosed after 250 mglkg/day. Spermatid and late spermatocyte popula groups (NTP 1993b). tions were absent by day 11 for those doses. Progressive and Nelson et al. ( 1989) reported that 2-Methoxyethanol adminis dose-related depletion of the spermatocytes and early spermatid tered in liquid diet increased embryonic intracellular pH in rats population occurred with continued dosing. Affected sperma during critical stages of organogenesis. Concentrations above tocytes had general cellular shrinkage, increased cytoplasmic 73 mglkg/day produced total embryomortality. At lower doses, eosinophilia, and nuclear pyknosis. Mitochondrial swelling and cardiovascular malformations were observed. The teratogenic disruption, cytoplasmic vacuolation, and early condensation of NOAEL was 16 mglkg/day 2-Methoxyethanol. Approximately nuclear chromatin were detected. The NOAELs for this study 36 mg/kg/day caused 50% embryomortality and decreased the were 50 and 250 mglkg/day for 2-Methoxyethanol and 2-Etho survivor's performance in subsequent maze testing (Nelson et al. xyethanol, respectively. Rats treated with 250 or 500 mglkg/day 1989). for 2-Methoxyethanol had dose-related decreased testicular and prostate weights. Testicular damage was generally reversible Inhalation Toxicity within one full maturation cycle, but totally atrophic tubules The toxicity of glycol ethers for dividing spermatocytes ulti were observed in the testes of several rats, indicating a loss of mately decreased testicular weights and size, caused degenera spermatogonia. Doses of 500 or 1000 mglkg/day 2-Ethoxyetha tion of germinal epithelium of the testes, and increased flaccid nol produced degeneration of a severity similar to that caused ity of the testes in rats and rabbits that had inhaled 30, 100, or by 100 mglkg/day 2-Methoxyethanol. The degree of degenera 300 ppm (0.09, 0.31, or 0.93 mg/m3) 2-Methoxyethanol6 hours tion and depletion of the spermatocyte population was similar daily 5 days/week for 13 weeks (Miller et al. 1983; Rao et al. for 500 and 1000 mglkg/day 2-Ethoxyethanol, but onset was 1983). more rapid at the lower dose (Foster et al. 1983). Similar effects In two of five male New Zealand white rabbits, inhalation of were seen when primary mixed cultures of Sertoli and germ cells 30 ppm 2-Methoxyethanol6 hours per day, 5 days per week, for from the testes of Sprague-Dawley rats were exposed to 2 and 13 weeks caused a small to moderate reduction in testes size. 10 mM concentrations of methoxy- and ethoxyacetic acid, res One of five rabbits had degeneration of the testicular germi pectively; up to 50 mM 2-Methoxyethanol or 2-Ethoxyethanol nal epithelium and few spermatozoa. In the same study, a dose did not induce degenerative changes (Gray et al. 1985). of 100 ppm 2-Methoxyethanol caused similar, but more severe An NTP study (1993b) of intraperitoneal (i.p.) and oral ad effects in four of five (testes size reduction) and three of five ministration of undiluted 0.165 and 0.330 mllkg 2-Methoxyetha rabbits (degeneration). Some tubules were normal in appear nol to pregnant Wistarratsresultedin 65% (i.p.) and53.8% (oral) ance, whereas others had no germinal elements. Small, flaccid total embryotoxicity 8:t 0.165 mllkg and 100% embryotoxicity testes and diffuse, severe degenerative changes were observed for both dose routes at 0.330 ml/kg. The control group had 4.4% in the three surviving rabbits in the high-dose group. In male total embryotoxicity, 190 total implants, 4.4% were dead andre- Sprague-Dawley rats used in the same study, reduced testes size Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 61

TABLE6 Dose-finding teratology study-Sprague-Dawley rats (Nelson et al. 1981)

Concentration 2-Ethoxyethanol Dams killed Dams delivered Dams killed Dams delivered Day (ppm) (GD7-13) (GD 7-13) (GD 14-20) (GD 14-20) delivered

1200 46 resorptions (4),a Not evaluated Not evaluated 8 dead pupsb (2), 24 0 live fetuses 20 resorption sites 900 40 resorptions (3), Not evaluated 1 dead fetus (1 ), 31 dead pups (3), 24 0 live fetuses 11 live fetuses 36 sites 600 54 resorptions (6), 12 dead pups (11), 3 dead fetuses (2), 29 dead pups (6),c 24 19live fetuses 0 live pups 18live fetuses 62 sites 300 16live fetuses (2), 5 dead pups (3),d 1 dead fetus (1), 11 dead pups (4),d 22.5 0 dead or resorbed 32live pups 3 live fetuses 27live pups 200 Not evaluated Not evaluated Not evaluated 21132 pups 22.5 survived (4) "Number of dams in parentheses. hDead within 24 hours of delivery. cone dam kept five pups alive for 10 days. aThese litters culled to eight pups/litter, and 20/24 pups survived. GD, gestational day. and treatment-related microscopic lesions only occurred at the or 14-20 in a preliminary dose-finding study. The dams were highest dose (300 ppm). Moderate to severe degeneration of the then either killed on GD 20 or allowed to deliver and rear sur germinal epithelium in the seminiferous tubules was observed. viving offspring. The only effect of treatment was slightly pro Reduced numbers of spermatozoa and degenerating spermato longed gestation (approximately 48 hours) in the dams exposed zoa were seen in the epididymides of all affected rats. A slight on GD 14-20. Dams whose pups died were killed and implan reduction in the amount of prostatic secretory material was also tation sites were counted. (Results are found in Table 6.) In the detected (Miller et al. 1983). second part of this experiment, rats were again exposed on GD Rao et al. (1983) also observed reduced testes size and at 7-13 and 14-20. The treatment dose was 100 ppm. The number rophic seminiferous tubules in Sprague-Dawley rats exposed of surviving pups was not given, but gestation was increased by to 300 ppm 2-Methoxyethanol vapor 6 h/day, 5 days/week for approximately 12 hours in treated dams. Offspring were sub 13 weeks. Male and female rats (20-30 per group) inhaled jected to behavioral testing, which included rotorod testing for 30-300 ppm 2-Methoxyethanol vapors as above. Males were neuromuscular ability, avoidance and operant conditioning, ac bred to unexposed females to determine reproductive capability tivity wheel testing, and other tasks. Groups of offspring were and dominant lethality. Females were bred to unexposed males. also used for neurochemical analysis. In general, offspring of Exposure to 300 ppm 2-Methoxyethanol was enough to com treated rats were less active and had impaired performance in be pletely suppress fertility in males. Fertility was partially restored havioral testing. Acetylcholine, norepinephrine, dopamine, and by 13 weeks after treatment. Body weights decreased at this con protein levels were altered in the brains of the offspring. Greater centration. No dominant lethal effect or impaired fertility was differences between the treatment groups and controls occurred seen in males exposed to 30 or 100 ppm. No adverse reproduc in offspring of rats exposed to 2-Ethoxyethanol earlier ingesta tive effects were observed in females at any dose. The NOAEL tion (Nelson et al. 1981). for fertility and reproduction was 100 ppm 2-Methoxyethanol In later studies, it was observed that either paternally or ma (Rao et al. 1983). ternally exposed rats did poorly in behavioral testing compared McGregor et al. ( 1983) found that sperm abnormalities in CD to controls and had neurochemical deviations when 25 ppm 2- rats and B6C3F1 mice caused by inhalation of 500 ppm (7 h/day Methoxyethanol was inhaled 7 h/day, 7 days/week for 6 weeks for 1 or 5 days) before mating increased the number of pre (Nelson et al. 1984). and post-implantation losses and decreased the pregnancy rate. Exposure to 25 ppm did not affect reproduction. No signs of clinical toxicity were observed in exposed rats and mice other Dermal Toxicity than the failure to increase body weight during the testing period Reproductive and developmental toxicity also occur after der (McGregor et al. 1983) .. mal application of glycol ethers. In a Chernoff-Kavlock assay, Pregnant Sprague-Dawley rats were treated with 200- pregnant Alpk/AP rats were given 10 mllkg/day of 3, 10, 30, or 1200 ppm 2-Ethoxyethanol by inhalation 7 h/day on GD 7-13 100% 2-Methoxyethanol in physiological saline on GD 6-17. Distributed for comment only -- do not cite or quote

62 COSMETIC INGREDIENT REVIEW

TABLE 7 Chernoff-Kavlock assay-dermal application, rats (Wickramaratne 1986)

% 2-Methoxyethanol

Reproduction Saline parameters control 3% 10% 30% 100% No. dams with litter 10/10 10/10 6/10 0/10 All dams died Group mean weight change (g) ana 1-21 115.3 ± 12.9 113.3 ± 18.4 102 ± 13.07 No litters produced All dams died GD 7-17 80.6 ± 13.4 81.0 ± 16.8 68.3 ±9.2 No litters produced All dams died No viable litters Day 1b 10110 10110 516 No litters produced All dams died DayS 10/10 10/10 3/6 No litters produced All dams died Total litter size (live + dead) 116 112 30 No litters produced All dams died Mean no. live pups/litter Day 1 11.5 ± 2.84 11.2 ± 2.49 6.0±4.2 No litters produced All dams died DayS 11.3 ± 2.71 11.2 ± 2.49 4.0 ± 4.43 No litters produced All dams died %Survival (days 1-5) 98.5 100 43.75 No litters produced All dams died Mean live pup weight (g) Day 1 6.05 ±0.9 6.33 ± 0.55 5.83 ± 0.55 No litters produced All dams died DayS 9.61 ± 1.01 9.82 ± 1.55 10±0.99 No litters produced All dams died

a Gestational day. bPostpartum day.

The test substance was applied to a 5 cm2 area of shaved infras Hardin, Goad, and Burg (1984) applied 2-Ethoxyethanol and capular skin for 6 hours per application. The site was covered other ethylene glycol ethers in water to the skin of time-mated by an occlusive patch. Control rats received saline. Rats were Sprague-Dawley rats on GD 7-16 and observed no toxic signs allowed to litter normally and rear their litters until day five post in the dams after treatment with 2-Ethoxyethanol. A dose vol partum. No adverse effects were observed after treatment with ume of 0.25 rnl 2-Ethoxyethanol was applied four times daily 3% 2-Methoxyethanol. Materno- and embryotoxicity (Table 7) to 18 rats. This dose was chosen as a teratogenic dose based on were observed at the higher doses (Wickramaratne 1986). the results of a previous study and served as the positive control Upon administration of a single bolus dose of 250 mglkg 2- for the remaining ethers tested. Dams were killed on GD 21. Methoxyethanol to CD-1 mice on GD 11, the ether was rapidly Dams treated with 2-Ethoxyethanol did have significant body oxidized to methoxyacetic acid. Seventy percent of the fetuses weight loss from GD 7-21 and the gravid uterus weights were had nonspecified digit anomalies. In order to mimic human der reduced also. Vehicle control results are in parentheses. Seven mal exposure to 2-Methoxyethanol, osmotic mini pumps were of 18 rats had totally resorbed litters (none). The number of im implanted subcutaneously in pregnant mice on GD 11. Each plants per litter did not differ from the control. The number of pump delivered 27 mg/kglh of the test chemical for 8 or 12 h. Ma dead implants per litter was 7.7 ± 5.1 (1.2 ± 1.2). In each litter ternal plasma, embryos, and exocoelomic/arnniotic fluid were with live fetuses, the number of live fetuses was 6.5 ± 4.1 for collected at 4, 8, 12, and 24 hours. When one pump was im 11litters (10.4 ± 2.6 for 17litters). The mean fetal body weight planted for 8 or 12 hours, 31% or 35% of fetuses examined was 3.0 ± 0.6 g (3.9 ± 0.6 g). Three fetuses in three litters had at GD 18 had digit malformations (micro-, syn-, ectro-, and gross malformations (acaudia and imperforate anus), whereas polydactyly, respectively). With two pumps, the percentage in none were found in the control group. Visceral abnormalities in creased to 87% or 85%. Peak concentrations of the metabo cluded cardiovascular malformations (aorta, ductus, ventricular lite, methoxyacetic acid, were detected 12 hours after exposure. septal defects, and abnormal subclavian), renal anomalies (hy Concentrations of 2.1 (one pump) or 5.2 (two pumps) mmollkg dronephrosis, hydroureter, ectopic kidney, and fused kidney), methoxyacetic acid were found in collected embryos, but ap hydrocephalus and hemorrhage of the brain, ocular defects (mi proximately 25% less was found in maternal plasma and >50% croophthalmia, abnormal optic nerve, and folded retina), and tes more in the exocoelomic/arnniotic fluid (Clarke, Duignan, and ticular abnormalites (slight ectopic or undescended testes, and Welsch 1990). agenesis). The control group had no cardiovascular, neural, or Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 63 ocular malformations, but had some instances of hydronephro were fertilized in vitro. Samples of zygotes were incubated for sis, and ectopic kidneys. The incidences were 10 of 10 litters 20 minutes (at 22°C} with 1.5, 3.0, and 6.0 M Propylene Glycol (21 fetuses of 34) with malformations in the 2-Ethoxyethanol in phosphate-buffered saline. There were three zygote cultures treated group, and 3 of 17 litters (3 fetuses of 87) in the vehicle per test concentration of Propylene Glycol, and each group of control group (Hardin, Goad, and Burg 1984). three was incubated with 0, 0.1, and 0.25 M sucrose, respec tively. The three control cultures without Propylene Glycol were also incubated with 0, 0.1, and 0.25 M sucrose, respectively. PROPYLENE GLYCOL Subsequently, the zygotes were incubated for 24 hours (at 37° C) In order to understand the structure activity relationship be under 5% C02, and the percentage of zygotes that cleaved to tween Ethylene Glycol and Propylene Glycol, the CIR Expert form two-cell embryos was determined. The percentage of zy Panel reviewed reproductive and developmental toxicity data on gotes that developed to two-cell embryos was not altered in the latter chemical. Propylene Glycol and Polypropylene Glycol control cultures or cultures exposed to 1.5 M Propylene Glycol have previously been reviewed by the CIR Expert Panel, and in (78% in both cultures), but was reduced in cultures exposed to formation from that safety assessment is presented in italicized 3.0M Propylene Glycol (7%; p < .05). Embryonic development text below (Andersen 1994). was inhibited completely in zygotes exposed to 6.0 M Propylene A continuous breeding reproduction study was conducted Glycol. The presence of sucrose in the incubation medium did using COBS Crl:CD-1 (ICR)BR outbred Swiss albino mice (6 not influence embryonic development (Damien, Luciano, and weeks old). The continuous breeding phase of the study (task II) Peluso 1989). In a later publication by the same investigators, was begun after the dose-setting study (task I) and involved three the data indicated that the exposure of B6D2F1 mouse zygotes experimental groups (40 mice per group) and a control group of to ~2.5 M Propylene Glycol for 2-7 minutes altered both in 80 mice. Experimental and control groups contained an equal tracellular pH and developmental potential. In that these effects number ofmale and female mice. The three experimental groups were independent ofvolume changes noted in zygotes and there were given the following doses (in feed or water), respectively, fore intracellular Propylene Glycol concentrations, the authors during a 7-day premating period: 1. 0% Propylene Glycol (daily postulated that the toxicity of Propylene Glycol is mediated by dose 1.82 glkg), 2.5% Propylene Glycol (daily dose 4.80 g!kg), direct alteration of the cell membrane (Damien, Luciano, and and 5.0% Propylene Glycol (daily dose 10.10 g!kg). The ani Peluso 1990). mals were then randomly grouped as mating pairs, cohabited, Kavlock, Short, and Chernoff (1987) investigated the ter and treated continuously for 98 days; data were collected on atogenic potential of several substances, including Propylene all newborns. If significant adverse effects on fertility were ob Glycol. A group of 30 pregnant female CD-1 mice was given served, a crossover mating trial (task Ill) was usually peiformed single oral doses of 10,000 ppm Propylene Glycol on days 8-12 to determine whether Fo males or females were more sensitive of gestation. Fertility rates, numbers of maternal deaths, num to the effects. Task III was not conducted and would have con bers of resorptions, average litter sizes, birth weights, and pup sisted ofmating high-dose mice ofeach sex to control mice ofthe postnatal weight gain were monitored in an assessment of the opposite sex and then analyzing the offspring. To peiform an off maternal and perinatal effects of Propylene Glycol. The fertil spring assessment of reproductive function (task N) following ity rates of mice given Propylene Glycol were not significantly exposure to Propylene Glycol, the dam (from phase II) was dosed different from those of control mice. There were no maternal through weaning and F1 mice were dosed until mating occurred deaths or resorptions observed in any of the animals dosed with at 74 ± 10 days of age. Mating pairs consisted of male and fe Propylene Glycol. All other parameters measured were not sig male offspring from the same treatment group (20/group/sex); F2 nificantly different from control values. Propylene Glycol was litters were examined. In the continuous breeding phase (task not a teratogen in this test. II), there were no significant changes (p < .05) in mean live Propylene Glycol (dose 0.01 mVg body weight) was also in pup weight per litter between the control group and any of the jected subcutaneously into each of21 pregnant ICR!Jclfemale treatment groups. In task N, which was an offspring assessment mice (9-12 weeks old) on day 9, 10, or 11 of gestation. Days of reproductive function, only the high-dose group (5% Propy 9-11 correspond to the sensitive stage for the induction offetal lene Glycol) was involved. There were no significant differences deaths and malformations in this strain of mice. The following ( p < .05) between control and experimental groups with re malformations were noted in 5 of the 226 living fetuses: open spect to the following observations in task N: mating index, eyelid (3 fetuses), polydactyly (1 fetus), and cleft palate (1 fetus). fertility index, mean number of live pups per litter, proportion In a control group of 28 mice (same strain and weight range) ofpups born alive, and sex ofpups born alive (Morrissey et al. injected subcutaneously with water (0.01 mVg body weight) dur 1989). ing pregnancy, the only malformation noted was exencephalus in The effect ofPropylene Glycol on the development ofB6D2F1 1 of 320 living fetuses. The incidence ofmalformations in 1,026 mouse zygotes in the pronuclear stage was evaluated; oocytes living fetuses from an untreated control group of 90 pregnant Distributed for comment only -- do not cite or quote

64 COSMETIC INGREDIENT REVIEW mice (same strain and weight range) was 3 fetuses with poly OH-(CHm-0-(CHm-CH:s dactyly, exencephalus, and open eyelid, respectively (Nomura 1977). 2-butoxyethanol Ascitic mouse ovarian tumor cells were used in an in vitro teratogenicity assay by Braun et al. (1982). Tumor cells were labeled with [3H]thymidine in situ via an intraperitoneal in jection of 0.2 mCi of radioactivity. Cells were harvested and OH-(CHm-O-(CH2) -CH:s suspended in phosphate-buffered saline (10 7 cellslml). Various 15 concentrations of Propylene Glycol were added to aliquots of ceteth-1 the cells (all concentrations not stated) and incubated at 37'C for 30 minutes. The teratogenicity of the substances tested was assayed by the ability of the test substances to inhibit attach ment of the tumor cells to Concanavalin A-coated plastic. The extent of attachment of the cells was measured by counting the OH-(CH2)2-0-(CJ:12>a radioactivity on the Concanavalin A-coated plastic. Propylene I Glycol did not inhibit attachment of the tumor cells; the largest CH dose tested was 27,000 mg/L Propylene Glycol. Propylene Gly II col was a nonteratogen in this test. CH-(CH2)r-CH;s oleth-1 DISCUSSION It is generally recognized that the ethylene glycol monoalkyl ethers are not themselves the toxic agent for reproductive or developmental toxicity, but rather it is one or more metabolites (Foster et al. 1984; Gray et al. 1985; Chiewchanwit and Au 1994). The ethylene glycol monoalkyl ethers are metabolized ~ R--C-O-CH2-CH-C~-(OCH~H2)n-OH by alcohol dehydrogenase and aldehyde dehydrogenase, result I ing in 2-methoxyethanol, for example, then the metabolites are OH methoxyacetaldehye and methoxyacetic acid. It is also appar ent from the available data that the toxicity of the metabolites of ethylene glycol monoalkyl ethers is inversely proportional PEG-n glyceryl cocoate (R is mainly C18) to the length of the alkyl chain, with methoxyacetic acid > FIGURE 3 ethoxyacetic acid > propoxyacetic acid > butoxyacetic acid Structures of ceteth-1, oleth-1, and PEG-n glyceryl cocoate (Gray et al. 1985). The CIR Expert Panel noted adverse re compared to 2-butoxyethanol. productive and developmental effects of oral administration of 2-butoxyethanol and the absence of any reproductive toxicity Implications for Oleths when applied dermally in its earlier review, concluding that 2- butoxyethanol can be used safely (at concentrations up to 10% ). Given the methods of manufacture of the Oleths (CIR 1996b ), there is no likelihood of methoxyethanol, ethoxyethanol, etc., being present as an impurity. Because there likely would be Implications for Ceteths ethylene glycol monomer linked to oleyl alcohol in preparations Given the methods of manufacture of the Ceteths (CIR 1996a), of the shorter chain length Oleths, it is appropriate to evaluate the there is no likelihood of methoxyethanol, ethoxyethanol, etc., potential toxicity of Oleth-1. As shown in Figure 3, compared to being present as an impurity. Because there likely would be the structure of the least active ethylene glycol monoalkyl ether, ethylene glycol monomer linked to cetyl alcohol in preparations 2-Butoxyethanol, the structure of Oleth-1 has a very long alkyl of the shorter chain length Ceteths and Ceteth-1 itself is listed chain. as a cosmetic ingredient, it is appropriate to evaluate the po Because reproductive and developmental toxicity is inversely tential toxicity of Ceteth-1. As shown in Figure 3, compared to related to the chain length of the alkyl group, the Oleths would the structure of the least active ethylene glycol monoalkyl ether, appear to present even less risk than 2-Butoxyethanol. In addi 2-Butoxyethanol, the structure of Ceteth-1 has a very long alkyl tion, many of the Oleths will contain only a polyethylene glycol chain. base, further reducing the potential for any adverse effects of the Because reproductive and developmental toxicity is inversely kind seen for ethylene glycol monoalkyl ethers. related to the chain length of the alkyl group, the Ceteths would appear to present even less risk than 2-Butoxyethanol. In addi Implications for PEGs Cocamine tion, many of the Ceteths will contain only a polyethylene glycol Given the methods of manufacture of the PEGs Cocamine base, further reducing the potential for any adverse effects of the (CIR 1996c), there is no likelihood of methoxyethanol, ethoxy kind seen for ethylene glycol monoalkyl ethers. ethanol, etc., being present as an impurity. In particular, the data Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 65 presented above relates exposures of ethylene glycol monoalkyl also unlikely that the Lanolin moieties would be metabolized ethers to subsequent reproductive and developmental toxicity. (e.g., via tl-oxidation) to simple methyl, ethyl, etc., alkyl groups. PEG Cocamines are polyethylene glycol ethers of the primary In addition, most of the polyethylene glycol chain lengths used aliphatic amine derived from coconut oil, and are chemically dif in making the various PEGs Lanolin are 10 or longer, suggesting ferent from alkyl ethers. This suggests that the toxicity concerns that there would be very little monomer linked by an ether group would not exist for the PEGs Cocamine. to the lanolin moiety.

Implications for PEGs Distearate Implications for PEGs Soya Sterol Given the methods of manufacture of the PEGs Distearate Given the methods of manufacture of the PEGs Soya Sterol (CIR 1996d), there is no likelihood of methoxyethanol, ethoxye (CIR 1996g), there is no likelihood of methoxyethanol, ethoxye thanol, etc., being present as an impurity. In particular, the data thanol, etc., being present as an impurity. In particular, PEG-5, presented above relate exposures of ethylene glycol monoalkyl -10, -16, -25, and -40 Soya Sterol would not be expected to be re ethers to subsequent reproductive and developmental toxicity. productive toxins as they are ethers of soybean oil sterols such as PEG Distearates are diesters of polyethylene glycol, and are stigmasterol, y-sitosterol, and campesterol, and are chemically chemically different from alkyl ethers. This suggests that the different from alkyl ethers. They have a polyethylene base, not toxicity concerns would not exist for the PEGs Distearate. ethylene glycol. In addition, most of the polyethylene glycol chain lengths used in making the various PEGs Soya Sterol are Implications for PEGs Glyceryl Cocoate 10 or longer, suggesting that there would be very little monomer Given the methods of manufacture of the PEGs Glyceryl linked by an ether group to the soybean oil sterols, further re Cocoate (CIR 1996e), there is no likelihood ofmethoxyethanol, ducing the potential for any adverse effects of the kind seen for ethoxyethanol, etc., being present as an impurity. In addition, the ethylene glycol monoalkyl ethers. PEGs Glyceryl Cocoate, the partial esters of polyoxyethylated glycerol with stearic acid, generally conform to the formula in CONCLUSION Figure 3. Metabolites of ethylene glycol monoalkyl ethers are repro where RCO- represents the coconut oil-derived fatty acids ductive and developmental toxins. In general, these metabolites (mainly C1s) and n has an average value equal to the number in of concern are not expected to be formed in cosmetic formula the name. Prior to the reaction with the RCO- group, polyethy tions that contain polymers of ethylene glycol. lene glycol is combined via an ether linkage with glycerol. Al though the PEGs Glyceryl Cocoate each have a polyethylene glycol base, not ethylene glycol, it should be recognized that REFERENCES ethylene glycol monomer could be present in any polymeriza American Conference of Governmental Industrial Hygienists (ACGIH). 1986. tion process that seeks to create chain lengths of less than 10. Threshold limit values and biological exposure indices for 1986-1987, 11, For that reason it is appropriate to evaluate the potential toxicity 18-19,23. Cincinnati, OH: ACGIH. Andersen, F. A. 1994. Final report on the safety assessment of Propylene Glycol of PEG-1 Glyceryl Cocoate. As shown in Figure 3, compared to and Polypropylene Glycols.l Am. Coll. Toxicol. 13:437-491. the structure of the least active ethylene glycol monoalkyl ether, Andersen, F. A. 1996. Final report of the safety assessment of Butoxyethanol. 2-Butoxyethanol, the structure of PEG-1 Glyceryl Cocoate has J. Am. Coll. Toxicol. 15:462-526. a very long complex chain that includes an ester linkage and an Beattie, P. J., and M. J. Brabec. 1986. Methoxyacetic acid and ethoxyacetic acid hydroxyl group. Because reproductive and developmental tox inhibit mitochondrial function in vitro. J. Biochem. Toxicol. 1:61-70. Brann, A. G., C. A. Buckner, D. J. Emerson, and B. B. Nichinson. 1982. Quanti icity is inversely related to the chain length of a simple alkyl tative correspondence between the in vivo and in vitro activity of teratogenic group, the PEGs Glyceryl Cocoate would appear to present no agents. Proc. Natl. Acad. Sci. USA 79:2056-2060. such risk. In addition, the PEGs Glyceryl Cocoate would contain Carney, E. W. 1994. An integrated perspective on the developmental toxicity of mostly a polyethylene glycol base, further reducing the poten ethylene glycol. Reprod. Toxicol. 8:99-113. tial for any adverse effects of the kind seen for ethylene glycol Carney, E. W., A. B. Liberacki, M. J. Bartels, and W. J. Breslin. 1996. Identifi cation of proximate toxicant for ethylene glycol developmental toxicity using monoalkyl ethers. rat whole embryo culture. Teratology 53:38-46. Chapin, R. E., and J. C. Lamb. 1984. Effects of ethylene glycol monomethyl Implications for PEGs Lanolin ether on various parameters of testicular function in the F344 rat. Environ Given the methods of manufacture of the PEGs Lanolin (CIR Health Perspect. 57:219-224. Chapin, R. E., S. L. Dutton, M.D. Ross, andJ. C. Lamb. 1985. Effects of ethylene 1996f), there is no likelihood of methoxyethanol, ethoxyethanol, glycol monomethyl ether (EGME) on mating performance and epididymal etc., being present as an impurity. Although the exact structures sperm parameters in F344 rats. Fundam. Appl. Toxicol. 5: 182-189. of lanolin, hydrogenated lanolin, lanolin oil, or lanolin wax are Cheever, K. L., H. B. Plotnik, D. E. Richards, and W. W. Weigel. 1984. not known, such extracts are usually long-chain compounds. Metabolism and excretion of 2-Ethoxyethanol in the adult male rat. Envi When combined via an ether linkage with polyethylene glycol, ron. Health Perspect. 57:241-248. Chiewchanwit, T., and W. W. Au. 1994. Cytogenic effects of2-methoxyethanol it is not likely that any of them would present the simple R and its metabolite, methoxyacetaldehyde, in mammalian cells in vitro. Mutat. group appearance of methyl, ethyl, propyl, or even butyl. It is Res. 320:125-132. Distributed for comment only -- do not cite or quote

66 COSMETIC INGREDIENT REVIEW

Clarke, D. 0., J. M. Duignan, and F. Welsch. 1990. Embryo dosimetry and Miller, R. R. 1987. Metabolism and disposition of glycol ethers. Drug Metab. incidence of malformations in CD-1 mice following subcutaneous infusion Rev. 18:1-22. of 2-methoxyethanol (2-ME). Teratology 41:544. Miller, R. R., J. A. Ayres, J. T. Young, and M. J. McKenna. 1983. Ethylene Cosmetic Ingredient Review (CIR). 1996a. Final report ofthe safety assessment glycol monomethyl ether. I. Subchronic vapor inhalation study with rats and ofCeteth-1, -2, -3, -4, -5, -6, -10, -12, -14, -15, -16, -20, -24, -25, -30, and rabbits. Fundam. Appl. Toxicol. 3:49-54. -45. Washington, DC: CIR. Miller, R. R., E. A. Hermann, J. T. Young, T. D. Laundry, and L. L. Calhoun. CIR. 1996b. Final report of the safety assessment of Oleth-2, -3, -4, -5, -6, -7, 1984. Ethylene glycol monomethyl ether and propylene glycol monomethyl -8, -9, -10, 12, -15, -16, -20, -23, -25, -30, -40, -44, and -50. Washington, ether: Metabolism, disposition, and subchronic inhalation toxicity studies. DC: CIR.2 Environ. Health Perspect. 57:233-239. CIR. 1996c. Final report of the safety assessment of PEG-2, -3, -5, -10, -15, Morrissey, R. E., J. C. Lamb, R. W. Morris, R. E. Chapin, D. Y. Gulati, and and -20 Cocamine. Washington, DC: CIR.2 J. J. Heindel. 1989. Results and evaluations of 48 continuous breeding CIR. 1996d. Final report of the safety assessment of PEG-2, -3, -4, -8, -9, -12, reproduction studies conducted in mice. Fundam. Appl. Toxicol. 13:747-777. -20, -32, -50, -75, -120, -150, and -175 Distearate. Washington, DC: CIR.2 National Center for Toxicological Research (NCTR). 1984a. Teratogenic CIR. 1996e. Draft report of the safety assessment of PEG-7 Glyceryl Cocoate. evaluation of Ethylene Glycol (CAS No. 107-21-1) administered to CD-1 Washington, DC: CIR.2 mice of gestational days 6 through 15. NTIS No. PB85-105385. CIR. 1996f. Final report of the safety assessment of PEG-5, -10, -20, -24, -25, NCTR. 1984b. Teratologic evaluation of Ethylene Glycol (CAS No. 107-21-1) -27, -30, -35, -40, -50, -55, -60, -75, -85, and -100 Lanolin; PEG-5, -10, -20, administered to CD rats on gestational days 6 through 15. NTIS No. -24, -30, and -70 Hydrogenated Lanolin; PEG-75 Lanolin Oil; and PEG-75 PB85-104594. Lanolin Wax. Washington, DC: CIR.2 National Institute for Occupational Safety and Health (NIOSH). 1983. Current CIR. 1996g. Draft report of the safety assessment ofPEG-5, -10, -16, -25, and Intelligence Bulletin Number 39. Glycol Ethers, 2-Methoxyethanol and -40 Soy Sterol. Washington, DC: CIR.2 2-Ethoxyethanol. NTIS No. PB84-155142. Creasy, D. M., J. C. Flynn, T. J. Gray, and W. H. Butler. 1985. A quantitative National Toxicology Program (NTP). 1991. Final report on the developmental study of stage-specific spermatocyte damage following administration of toxicity of Ethylene Glycol (CAS No. 107-21-1) in New Zealand white ethylene glycol monomethyl ether in the rat. Exp. Mol. Pathol. 43:321-336. rabbits-Volume 1 of 2: Final study report and appendix. NTIS No. Damian, M., A. A. Luciano, and J. J. Peluso. 1989. Propanediol-induced PB91-211219. alterations in membrane integrity, metabolism, and developmental potential NTP. 1993a. Toxicology and carcinogenesis studies of Ethylene Glycol (CAS of mouse zygotes. Hum. Reprod. 4:969-974. No. 107-21-1) in B6C3FJ mice (feed studies). NTP Technical Report Series Damian, M., A. A. Luciano, and J. J. Peluso. 1990. Propanediol alters intra No. 413, NIH Publication No. 93-3144. Research Triangle Park, NC: NTP. cellular pH and developmental potential of mouse zygotes independently of NTP. 1993b. Technical report on toxicity studies of ethylene glycol ethers volume change. Hum. Reprod. 5:212-216. 2-Methoxyethanol, 2-Ethoxyethanol, 2-Butoxyethanol (CAS Nos. 109-86-4, Foster, P.M. D., D. M. Creasy, J. R. Foster, L. V. Thomas, M. W. Cook, and 110-80-5, 111-76-2) administered in drinking water to F344/N rats and S. D. Gangolli. 1983. Testicular toxicity of ethylene glycol monomethyl and B6C3FJ mice. NTP Toxicity Report Series No. 26, NIH Publication No. monoethyl ethers in the rat. Toxicol. Appl. Pharmacol. 69:385-399. 93-3349. Research Triangle Park, NC: NTP. Foster, P.M., D. M. Creasy, J. R. Foster, and T. J. Gray. 1984. Testicular toxicity Neeper-Bradley, T. L., R. W. Tyl, L. C. Fisher, M. F. Kubena, M. A. Vrbanic, produced by ethylene glycol monomethyl and monoethyl ethers in the rat. and P. E. Losco. 1995. Determination of a no-observed-effect level for Environ. Health Perspect. 57:207-217. developmental toxicity of ethylene glycol administered by gavage to CD rats Gray, J. B., E. J. Moss, D. M. Creasy, and S.D. Gangolli. 1985. Studies on the and CD-1 mice. Fundam. Appl. Toxicol. 27:121-130. toxicity of some glycol ethers and alkoxyacetic acids in primary testicular Nelson, B. K., W. S. Brightwell, J. V. Setzer, B. J. Taylor, and R. W. Hornung. cell cultures. Toxicol. Appl. Pharmacal. 79:490-501. 1981. Ethoxyethanol behavioral teratology in rats. Neurotoxicology Hardin, B. D., P. T. Goad, and J. R. Burg. 1984. Developmental toxicity offour 2:231-249. glycol ethers applied cutaneously to rats. Environ. Health Perspect. 57:69-74. Nelson, B. K., W. S. Brightwell, J. R. Burg, and V. J. Massari. 1984. Behavioral Jacobsen, D., and K. E. McMartin. 1986. Methanol and ethylene glycol and neurochemical alterations in the offspring of rats after maternal or poisonings: Mechanism of toxicity, clinical course, diagnosis and treatment. paternal inhalation exposure to the industrial solvent 2-Methoxyethanol. Med. Toxicol. 1:309-344. Pharmacal. Biochem. Behav. 20:269-279. Kavlock, R. J., R. D. Short, and N. Chernoff. 1987. Further evaluation of an in Nelson, B. K., C. V. Vorhees, W. J. Scott, and L. Hastings. 1989. Effects of vivo teratology screen. Teratogen Carcinog. Mutagen 7:7-16. 2-Methoxyethanol on fetal development, postnatal behavior, and embryonic Kennedy, C. H., I. Y. Chang, and R. F. Henderson. 1993. Effect of dose on intracellular pH in rats. Neurotox. Teratol. 11:273-284. the disposition of 2-Ethoxyethanol after inhalation by F344/N rats. Fundam. Nomura, T. 1977. Similarity of the mechanism of chemical carcinogen-initiated Appl. Toxicol. 21:486-491. teratogenesis and carcinogenesis in mice. Cancer Res. 37:969-973. Khera, K. S. 1991. Chemically induced alterations in maternal homeostasis and Oudiz, D. J., K. Walsh, and L. M. Wiley. 1993. Ethylene glycol monomethyl histology of conceptus: Their etiologic significance in rat fetal anomalies. ether (EGME) exposure of male mice produces a decrease in cell proliferation Teratology 44:259-297. ofpreimplantation embryos. Reprod. Toxicol. 7:101-109. Lamb, J. C., R. R. Maronpot, D. K. Gulati, and V. S. Russell. 1985. Reproductive Price, C. J., C. A. Kimmel, R. W. Tyl, and M. C. Marr. 1985. The developmental and developmental toxicity of ethylene glycol in the mouse. Toxicol. Appl. toxicity of ethylene glycol in rats and mice. Toxicol. Appl. Pharmacal. Pharmacal. 81:100-112. 81:113-127. Maronpot, R. R., J.P. Zelenak, E. V. Weaver, and N.J. Smith. 1983. Terato Rao, K. S., S. R. Cobel-Geard, J. T. Young, T. R. Hanley, Jr., W. C. Hayes, genicity study of ethylene glycol in rats. Drug Chern. Toxicol. 6:579-594. J. A. John, and R. R. Miller. 1983. Ethylene glycol monomethyl ether. IT. McGregor, D. B., M. J. Willins, P. McDonald, M. Holstrom, D. McDonald, Reproductive and dominant lethal studies in rats. Fundam. Appl. Toxicol. 3: and R. W. Niemeier. 1983. Genetic effects of 2-methoxyethanol and 80-85. bis(2-methoxyethyl)ether. Toxicol. Appl. Pharmacal. 70:303-316. Ritter, E. J., W. J. Scott, Jr., J. L. Randall, and J. M. Ritter. 1985. Teratogenicity of dimethoxyethyl phthalate and its metabolites methoxyethanol and methoxyacetic acid in the rat. Teratology 32:25-31. Scott, W. J., R. Fradkin, W. Wittfoht, and H. Nau. 1989. Teratologic potential 2 Available for review: Director, Cosmetic Ingredient Review, 110117th Street, of 2-methoxyethanol and transplacental distribution of its metabolite, NW, Suite 310, Washington, DC 20036, USA. 2-methoxyacetic acid, in non-human primates. Teratology 39:363-373. Distributed for comment only -- do not cite or quote

ETHYLENE GLYCOL AND ITS ETHERS 67

Sun, J. D., S. W. Frantz, and J. L. Beskitt. 1995. In vitro skin penetration of Tyl, R. W., C. J. Price, M. C. Marr, and C. B. Myers. 1993. Developmental ethylene glycol using excised skin from mice and humans. J. Toxicol. Cut. toxicity evaluation of ethylene glycol by gavage in New Zealand White Ocular Toxicol. 14:273-286. rabbits. Fundam. Appl. Toxicol. 20:402-412. Tyl, R. W., L. C. Fisher, M. F. Kubena, M. A. Vrbanic, and P. E. Losco. Welsch, F., and D. B. Stedman. 1984a. Inhibition ofintercellularcommunication 1995a. Assessment of the developmental toxicity of ethylene glycol applied between normal human embryonal palatal mesenchyme cells by teratogenic cutaneously to CD-1 mice. Fundam. Appl. Toxicol. 27:155-166. glycol ethers. Environ. Health Perspect. 57:125-133. Tyl, R. W., B. Ballantyne, L. C. Fisher, D. L. Fait, D. E. Dodd, D. R. Klonne, Welsh, F., and D. B. Stedman. 1984b. Inhibition of metabolic cooperation I. M. Pritts, and P. E. Losco. 1995b. Evaluation of the developmental toxicity between Chinese Hamster V79 cells by structurally diverse teratogens. of ethylene glycol aerosol in the CD rat and CD-1 mouse by whole-body Teratog. Carcinog. Mutagen 4:285-301. exposure. Fundam. Appl. Toxicol. 27:57-75. Wickrarnaratne, G. A. de S. 1986. The teratogenic potential and dose-response Tyl, R. W., L. C. Fisher, M. F. Kubena, P. E. Losco, and M.A. Vrbanic. 1989. De of dermally administered ethylene glycol monomethyl ether (EGME) termination of a developmental toxicity "no observable effect level" (NOEL) estimated in rats with the Chernoff-Kavlock assay. J. Appl. Toxicol. 6:165- for ethylene glycol (EG) by gavage in Swiss mice. Teratology 39:487. 166. Distributed for comment only -- do not cite or quote

Distributed for comment only -- do not cite or quote

Final Report on the Safety Assessment of Propylene Glycol and Polypropylene Glycols 1

Abstract: Propylene Glycol is an aliphatic alcohol manufactured as a reaction product of propylene oxide and water. Polypropylene Glycol is a polymer formed by adding propylene oxide to dipropylene glycoL Propylene Glycol is reportedly used as a skin-conditioning agent-humectant, solvent, viscosity decreasing agent, and humectant in thousands of cosmetic formulations. Poly propylene Glycols of various polymer lengths are reportedly used as miscel laneous skin-conditioning agents in far fewer formulations. Acute, subchronic, and short-term animal studies suggested little toxicity beyond slight growth and body weight decreases. Little ocular or skin irritation was observed in animal studies, and no sensitization was seen. Small increases in fetal malformations were seen in mice injected subcutaneously with Propylene Glycol, but a con tinuous breeding reproduction study in mice showed no reproductive toxicity following oral administration. A wide range of mutagenesis assays were neg ative, and studies in mice and rats showed no evidence of carcinogenesis. Clinical data showed skin irritation and sensitization reactions in Propylene Glycol in normal subjects at concentrations as low as 10% under occlusive conditions and dermatitis patients at concentrations as low as 2%. A careful evaluation of skin irritation and sensitization data as a function of disease state of the individual, occlusion, and concentration was done. On the basis of that analysis, it is concluded that Propylene Glycol and Polypropylene Glycol are safe for use in cosmetic products at concentrations up to 50%. Key Words: Propylene Glycol-Polypropylene GlycoL

CHEMISTRY Definition and Structure

Propylene Glycol (PG) is an aliphatic alcohol that conforms to the following formula (Estrin et al., 1982a):

CH3CHCH20H I OH PG (CAS No. 57-55-6) is also known as 1,2-Propanediol, 1,2-Dihydroxy propane, Methylethylene Glycol, Methyl Glycol, Monopropylene Glycol, Pro-

1 Reviewed by the Cosmetic Ingredient Review Expert Panel. Address correspondence and reprint requests to Dr. F. A. Andersen at Cosmetic Ingredient Re view, 1101 17th Street NW, Suite 310, Washington, D.C. 20036, U.S.A.

437 Distributed for comment only -- do not cite or quote

438 COSMETIC INGREDIENT REVIEW pane-1,2-Glycol, a-Propyleneglycol, 1,2-Propylene Glycol, and numerous trade names (Estrin et al., 1982a; Sweet, 1987). Polypropylene Glycol (PPG) is a polymer of PG and water that generally con forms to the following formula (Estrin et al., 1982a; Food Chemicals Codex, 1981):

n ::::; number of oxypropylene groups Polypropylene Glycol (CAS No. 25322-69-4) is also known as Polyoxypropylene, PPG, and numerous trade names (Estrin et al., 1982a).

Properties The chemical and physical characteristics of PG and PPG are listed in Tables 1 and 2, respectively.

Analytical Methods PG has been analyzed via gas chromatography (Gebhardt, 1968; Demey et al., 1988) and ultraviolet (UV) spectral analysis (Dow Chemical Co., 1992; Eastman

TABLE 1. Chemical and physical properties of Propylene Glycol

Ref. Physical form Colorless, hygroscopic liquid Estrin et al. (1982b) Odor Practically odorless Sax (1979) Formula C3Hs02 Estrin et al. (1982a) M:olecular weight 76.11 Sweet (1987) Sax (1979) Assay 97.5% minimum (cosmetic use) Estrin et al. (1982b) 99.5% minimum (industrial use) Dow (1975, 1978) Vapor density 2.62 (air = 1) Sax (1979) Vapor pressure 0.08 mm Hg at 20°C Sax (1979) Refractive index 1.4324 at 20°C Sax (1979) Density 1.0361 at 25°/4°C Weast (1982) 1.0362 at 25°/25°C sax (1979) Specific gravity 1.035-1.037 at 25°/25°C Estrin et al. (1982b) Boiling ppint 189°C Weast (1982) 188.2°C Sax (1979) Freezing point -59°C Sax (1979) Flash point 210°F (open cup) Sax (1979) Autoignition temperature 700°F Sax (1979) Solubility Soluble in water, ethanol, benzene Weast (1982) Distillation range 184-189°C Estrin et al. (1982b) Acidity 0.002 mEq/ml maximum Estrin et al. (1982b) Sulfated ash 0.07% maximum Estrin et al. (1982b) Water 0.2% maximum Es~rin et al. (1982b) Arsenic (as As) 3 ppm maximum Estrin et al. (1982b) Explosion hazard Moderate when exposed to flame Sax (1979) Lower explosion limit 2.6% Sax (1979) Upper explosion limit 12.6% Fire hazard Low Sax (1979)

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 439

TABLE 2. Chemical and physical properties of Polypropylene Glycol

Ref. Physical form Clear, colorless, viscous liquid Food Chemicals Codex (1981) Formula HO(C3H60)nH Sax (1979) Molecular weight 400-2,000 Sax (1979) Density 1.002-1.007 Sax (1979) Flash point 390 to >440°F Sax (1979) Solubility Soluble in water, aliphatic ketones, and Food Chemicals Codex (1981) alcohols; insoluble in ether and aliphatic hydrocarbons Viscosity 85-97 centistokes at MW = 1 ,000 Food Chemicals Codex (1981) 150-175 centistokes at MW = 2,000 Food Chemicals Codex (1981) Arsenic (as As) 3 ppm maximum Food Chemicals Codex (1981) Heavy Metals (as Pb) 5 ppm maximum Food Chemicals Codex (1981) pH Between 6 and 9 Food Chemicals Codex (1981) Propylene oxide 0.02% maximum Food Chemicals Codex (1981) Residue on ignition 0.01% maximum Food Chemicals Codex (1981) Fire hazard Slight when exposed to heat or flame, Sax (1979) may react with oxidizing materials

Chemical Co., 1992; Istituto Ganassini spa di ricerche biochimiche, 1992). Each spectral analysis indicates no absorbance in the UV A and UVB bands. PPG 1200 has been detected at sub-ppm levels in aqueous and organic media by a method that combines the classic Zeisel alkoxy reaction with high-efficiency gas-liquid chromatography (Ramstad et al., 1978).

Methods of Manufacture All PG that is of commercial significance is manufactured as a reaction product of propylene oxide and water; no additives or solvents are used. Additionally, the chlorohydrin process is used to manufacture the reactant, propylene oxide, which is recovered as a pure product before conversion to glycol (Dow U.S.A., 1993). PG is also manufactured by treating propylene with chlorinated water to form the chlorohydrin; the chlorohydrin is converted to the glycol by treatment with sodium carbonate solution. Another method of preparation involves heating glyc erol with sodium hydroxide (Rothschild, 1988). PPGs are manufactured by the addition of propylene oxide to dipropylene gly col (Shaffer et al., 1951).

Impurities In cosmetic products, the purity of PG is specified as a minimum of 97 .5%. In industry, the purity ofPG is specified as a minimum of99.5%. Impurities found in PG may include sulfated ash, arsenic, propylene oxide, and heavy metals such as lead (Estrin, et al. 1982b; Food Chemicals Codex, 1981). The specific maximum levels of impurities that may be present in both food- and cosmetic-grade PG and PPG are listed in Tables 1 arid 2. The Dow Chemical Co. (Dow U.S.A.) recommends that USP-grade PG be used for cosmetics. USP-grade PG has a typical assay of 99.9% and a specification of

JAm Coil Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

..... ~ t 3 c g ::::: ~ ~ ~ :-.... .!"" TABLE 3. Product formulation data for Propylene Glycol (FDA, 1984) ~ (J ?- No. of product formulations within each concentration range (%) c .... ·IJ:l ~ Product category >50 25-50 10-25 5-10 1-5 0.1-1 0-0.1 Unknown Total ~ Baby shampoos 1 2 1 3 7 ~ Baby lotions/oils/powders/creams 1 5 2 8 (J Other baby products 2 1 3 Bath oils/tablets/salts 2 4 2 5 13 3 29 ~ Bubble baths 1 6 15 64 37 123 §5 Other bath preparations 2 1 1 14 5 20 5 48 Eyebrow pencil 1 1 ~ Eyeliner 2 7 42 4 55 t;j Eye shadow 3 71 18 44 39 175 <: Eye lotion 1 2 1 4 ~ ~ Eye makeup remover 2 2 1 8 4 17 l:ll Mascara 4 6 30 4 2 14 60 ~ Other eye makeup preparations 13 14 7 9 43 tti Colognes/toilet waters 5 8 27 16 40 96 Perfumes 1 3 4 7 13 28 ~ Powders dusting/talcum excl. aftershave talc 1 4 4 1 10 Sachets 4 6 14 2 2 28 Other fragrance preparations 1 4 9 19 4 6 43 Hair conditioners 7 17 10 10 14 58 Hair sprays (aerosol fixatives) 1 6 2 1 10 Hair straighteners 10 12 22 Permanent waves 3 2 25 2 11 43 Rinses (noncoloring) 1 3 1 5 3 13 Shampoos (noncoloring) 4 10 60 52 31 54 211 Tonics/dressings/other hair grooming aids 2 5 13 2 6 3 31 Distributed for comment only -- do not cite or quote

Wave sets 1 1 5 5 6 18 Other hair preparations 1 1 1 5 7 1 16 Hair dyes/colors (requiring caution statement) 60 140 82 1 5 288 Hair rinses (coloring) 4 18 7 29 Hair shampoos (coloring) 1 2 3 Hair lighteners with color 1 1 ~ Hair bleaches 2 4 6 Other hair coloring preparations 1 3 1 5 ~ Blushers (all types) 1 3 31 14 7 12 17 85 ~ Face powders 16 1 12 29 Foundations 45 106 44 1 10 56 262 ~ Leg and body paints 3 3 ~ Lipstick 1 3 244 389 544 ll81 Makeup bases 12 200 61 2 7 52 334 ~ Rouges 2 7 4 2 7 8 30 ~ Makeup fixatives 1 3 4 g Other makeup preparations 1 3 8 17 4 27 41 131 Cuticle softeners 6 3 1 2 12 ~ Nail creams/lotions 6 1 7 Nail polish and enamel removers 1 1 2 ~ Other manicuring preparations 1 2 3 6 ~ Dentifrices (aerosol/liquid/paste/powder) 1 1 2 g Mouthwashes/breath fresheners 3 3 Other oral hygiene products 1 1 ~ Bath soaps/detergents 5 6 4 24 39 ::tl Deodorants (underarm) 19 4 9 28 43 10 2 9 124 .... ~ ;... Douches 1 4 1 1 7 :! Feminine hygiene products 1 1 2 ~ 24 53 ~ Other personal cleanliness products 3 9 16 1 ~ :::: Aftershave lotions 1 5 49 28 8 6 97 ~ ~ ;:;· (Continued) ~ ~ ~ :-~ ..... g ~ V:i ~ 5>- ..... ~ ~ ...... Distributed for comment only -- do not cite or quote

..... ;... tN '" ~:::: ~ >< ;:;· TABLE 3. Continued ~ ~ No. of product formulations within each concentration range (%) ·"''- Product category >50 25-50 10--25 5-10 1-5 0.1-1 0--0.1 Unknown Total ~ Beard softeners 2 1 3 (J ?' Preshave lotions 1 1 2 4 8 a '- ~ '0 Shaving cream (aerosollbrushless/lather) 2 1 17 8 1 5 34 ~ Other shaving preparations 1 2 3 5 2 13 Skin cleansing products (cold creams/ ~ ~ lotions/liquids/pads) 2 15 67 128 12 23 29 276 (J Depilatories 2 2 2 6 Face/hand/body (excl. shaving preparations) 1 14 71 197 48 31 55 417 ~ Foot powders/sprays 1 1 Hormone products 1 1 2 1 5 sa Moisturizing products 2 5 99 170 18 40 24 358 gs Night preparations 1 6 30 50 7 6 5 105 til Paste masks (mud packs) 5 14 45 7 2 10 83 <; Skin lighteners 2 6 9 2 19 ....., Skin fresheners 1 10 56 17 15 37 136 ~ Wrinkle-smoothing products (removers) 1 2 6 2 2 1 14 ~ Other skin care preparations 1 4 25 55 18 14 32 149 Suntan gels/creams/liquids 2 14 30 3 12 15 76 :stl'j Indoor tanning preparations 1 9 2 12 ~ Other suntan preparations 1 4 5 1 4 15 Ingredient total 21 43 236 1,068 1,529 683 896 1,200 5,676 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 443

99.5% minimum purity. Dow U.S.A. recognizes that the United States Pharma copeia now allows up to 5 ppm propylene oxide in PG and is of the opinion that typical levels contained in products today are less than detectable amounts (Dow U.S.A., 1993).

COSMETIC USE PG is used as a skin-conditioning agent-humectant, solvent, viscosity decreasing agent, and humectant in cosmetic formulations (Nikitakis, 1988). Data on the uses of PG are summarized in Table 3. Data submitted to the Food and Drug Administration (FDA) in 1984 indicated that PG was used in a total of 5,676 cosmetic formulations. Use concentrations of PG in cosmetic formulations are listed in Table 3 as <0.1 to >50% (FDA, 1984). PPGs are used as miscellaneous skin-conditioning agents in cosmetic formula tions (Nikitakis, 1988) as listed in Tables~. Data on the uses of three of the PPGs, PPG 9, PPG 26, and PPG 425, indicate that PPG-9 has been used in a total of 6 cosmetic formulations at concentrations of 1-5 or 0.1-1% and PPG 26 in a total of 10 cosmetic formul~tions. The largest use of PPG 26 was in. blusher

products at a concentration of. 1-5%. Table 6 indicates that PPG 425. was used. in one cosmetic formulation, a hair bleach, at a concentration of between 1 and 5% (FDA, 1984). Volunt;:~.ry filing of product formulation data with the FDA by cosmetic manu facturers and formulators must conform to the format of concentration rang~s and product categories as described in Title 21, Part 720.4 of the Code of Federal Regulations (21 CFR 720.4). Since certain cosmetic ingredients are supplied to the formulator at <100% concentration, the concentration reported by the formulator may not necessarily reflect the actual concentration found in the finished cosmetic product; the actual concentration in such an instance would be a fraction of that reported to the FDA. The fact that data are submitted only within the framework of a "concentration range" provides the opportunity for overestimation of the actual concentration of an ingredient in a particular product. An entry at the lowest end of a concentration range is considered the same as one entered at the highest end of that range, thus introducing the possibility of a 2- to 10-fold error in the assumed ingredient concentration. Concentration of use data are no longer reported to FDA by the cosmetics industry (Federal Register, 1992). However, data on PG supplied to the Cosmetic, Toiletry, and Fragrance Association (CTF A) indicate that concentrations of use

TABLE 4. Product formulation data for PPG 9 (FDA, 1984)

No. of product formulations within each concentration range (%) . Product category 25-50 10-25 5-10 1-5 0.1-1 0-0.1 Unknown Total Other bath preparations 2 2 Shampoos (noncoloring) 2 2 4 Ingredient total 4 2 6

JAm Coli Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

444 COSMETIC INGREDIENT REVIEW

TABLE 5. Product formulation datafor PPG 26 (FDA, 1984)

No. of product formulations within each concentration range (%) Product category 25-50 10--25 5-10 1-5 0.1-1 0--0.1 Unknown Total Bath oils/tablets/salts 1 1 Blushers (all types) 3 3 Deodorants (underarm) 2 2 4 Moisturizing products 1 1 Other skin care preparations 1 1 Ingredient total 8 2 10 range between 3 and 5% in current products manufactured by one company. The intention of using PG in cosmetics at concentrations as great as 80% was also stated (CTF A, 1992). Current frequency of use data indicate that PG is used in a total of 4,056 cosmetic products. PPG 9 is used in 13 products, and PPG 12 and PPG 26 are used in 2 and 5 products, respectively (FDA, 1993). Products containing PG or PPG may contact the skin, hair, and eyes. Ingestion of small quantities of either compound is possible. Contact with either compound may be as little as once to as often as daily for several hours.

International Use PG, PPG 9, and PPG 26 are included in the list of ingredients that have been approved for use in cosmetic formulations marketed in Japan (Nikko Chemicals, Co., Ltd., 1991-92). PG and PPGs are not included in the list of ingredients that are prohibited from use in cosmetic products marketed in the European Economic Community (EEC Cosmetics Directive, 1990).

NONCOSMETIC USE PG is generally recognized as safe (GRAS) when used in accordance with good manufacturing practices (21 CFR 582.1666). A monograph prepared for the GRAS evaluation is available (Informatics, 1973). PG is also an accepted optional ingre dient in standardized food substances such as vanilla extract (21 CFR 169.175). In an assessment of food additives, the World Health Organization (WHO) recommended a maximum daily oral intake ofPG of25 mg/kg body wt/day (WHO, 1974). PG is an ingredient in Over-the Counter (OTC) drug products (FDA, 1991).

TABLE 6. Product formulation data for PPG 425 (FDA, 1984) No. of product formulations within each concentration range (%) Product category 25-50 10--25 5-10 1-5 0.1-1 0--0.1 Unknown Total Hair bleaches Ingredient total

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 445

FDA issued a proposed rule on the use of PG as a pediculicide, indicating that OTC drug products containing PG for this purpose are not generally recognized as safe and effective or are misbranded. An advance notice of proposed rulemaking containing a similar statement regarding use of PG as a hair grower was also issued by FDA. OTC drug products containing PG for use as a demulcent are generally recognized as safe and effective and are not misbranded; this statement constitutes final agency action (FDA, 1991). PG has been classified as both an excellent solvent and an effective preservative; it has replaced glycerin as the vehicle in many formulations of topical therapeutic agents (AMA Division of Drugs, 1983). Other documented noncosmetic uses of PG include use of the compound in the production of varnishes and resins and as an antifreeze (Angelini and Meneghini, 1981). PG has been used as a source of energy in animal diets (Emmanuel, 1976), as a cryoprotective agent in renal preservation (Pegg et al., 1987; Halasez and Collins, 1984), and as a therapy for ketosis in cattle (Ruddick, 1972). An investigation of PG as a potential environmental contaminant is available (National Technical Information Service, 1979). PPG is an indirect food additive that has been approved for use as a component of resinous and polymeric coatings (21 CFR 175.300).

GENERAL BIOLOGY Absorption, Metabolism, and Excretion Propylene Glycol In mammals, the major route of PG metabolism is to lactaldehyde and then lactate via hepatic alcohol and aldehyde dehydrogenases; in an alternative path way, lactaldehyde is metabolized to methylglyoxal before lactate is formed (Chris topher et al., 1990a). A study on the metabolism ofPG was conducted by Morshed et al. (1988). Five groups of six rats each were given oral doses of 4.83, 9.66, 19.32, 38.64, and 77.28 mmol of aqueous PG/kg body wt. Blood was collected from these animals at times ranging from 5 min to 24 h postdosing. Appearance of PG in the blood of the treated rats was in a dose-related manner. The maximum concentration of PG in the blood, 28 mmol/L, was found 2 h after dosing in the rats given the largest dose. In the second part of this experiment, the metabolism of PG was further inves tigated. The metabolism of PG in animals is initially catalyzed by several dehy drogenases. To test inhibition of PG metabolism, the authors pretreated five groups of six Wistar rats each with oral doses of 0.0, 0.025, 0.05, 0.20, and 1.0 mmol pyrazole, a known alcohol dehydrogenase inhibitor. The rats were then given oral doses of 4.83, 9.66, 19.32, 38.64, or 77.28 mmol/kg PG. Blood samples were collected in a manner similar to that in the preceding experiment. In the pretreated rats, the greatest blood concentrations of PG were found in the blood 30 min postdosing, regardless of the amount of compound administered. The peak concentration of PG in the blood of these animals was significantly reduced com pared with rats dosed with PG that were not pretreated with pyrazole.

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

446 COSMETIC INGREDIENT REVIEW

In a separate experiment, the same authors tested the urinary clearance of PG from rats pretreated with pyrazole. Three groups of rats, 36 per group, were given oral doses of 0.0, 0.20, or 1.0 mmol/kg pyrazole. Each group was subdivided into three subgroups (a total of nine groups) and given 19.32, 38.64, or 77.28 mmol/kg PG orally. The excretion of PG in the urine increased according to the amount of PG administered, both with and without pyrazole. The amount of PG excreted was significantly greater in the rats fed doses of pyrazole then in the control rats. The metabolism ofPG in the rabbit was investigated by Yu and Sawchuk (1987). Three groups of New Zealand White male rabbits, three animals per group, were given intravenous doses of0.5, 1.0, or 2.0 g/kg PG in normal saline solution. Blood and urine were collected between 0 and 12 h postdosing. The plasma concentra tion of PG was measured, and the half-lives of the compound in the blood were calculated. In rabbits given the smallest dose, the half-life was 64.0 min. For the 1.0- and 2.0-g/kg doses, the half-lives were 76.2 and 75.4 min, respectively. As the plasma concentration of PG diminished, the metabolism clearance of PG in creased proportionally. It was determined that there was a nonlinear relationship between the two routes of clearance of PG. After oral administration of PG (38.66 mmol/kg, single dose) to four adult New Zealand rabbits (weights 2-2.5 kg), the concentration of L-lactate in whole blood increased significantly (p < 0.01, n = 4) to 2.55 ± 0.62 mmol/L when compared with a fasting concentration of 1.04 ± 0.22 mmol/L. The concentration of o-lac tate was also significantly elevated (p < 0.001). o- and L-Lactate are PG metab olites. Blood concentrations of pyruvate remained mostly unaffected both before and after PG administration (Morshed et al., 1991). Speth et al. (1987) investigated the effects of PG administration using human subjects. Six patients with confirmed malignancies were given intravenous doses of a potentially cytostatic agent, mitoquidone, dissolved in PG. Subjects received doses of 5.1-21.0 g PG/day in a 4-h intravenous infusion. The concentration ofPG in the plasma and the half-life of PG in the body were measured. The average half-life was 2.3 ± 0.7 h, and the elimination of PG from the body was in a dose-related manner. The authors also measured lactic acid concentration, ve nous pH, plasma osmolality, hemoglobin, and haptoglobin levels in each test subject. No significant alterations in any of these parameters were noted.

Polypropylene Glycols

PPG 425 and PPG 1025 are readily absorbed from the gastrointestinal (GI) tract. Following administration of a single oral dose ofPPG 425 (0.5-1.0 g/kg) to rabbits, 17-39% was recovered from the urine. Under similar conditions of exposure, ~40% of administered PPG 1025 was excreted in the urine of rabbits. After rabbits received a single oral dose ofPPG 2025 (2 g/kg), 40-70% was excreted in the feces and a negligible amount was excreted in the urine. For all of the groups tested, neither the weight range, strain, nor number of animals involved was stated (Shaf fer et al., 1951). Dermal absorption ofPPG 425 has been demonstrated in rabbits (Shaffer et al., 1951) (see Acute Dermal Toxicity).

JAm Coli Toxicol, Vol. 13, No.6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 447

Cellular Effects The effects of PG on mouse urinary bladder epithelial cell growth were inves tigated by Farsund (1981). A group of hairless male Oslo strain mice (number not specified) was given subcutaneous injections of 0.2 ml PG 3 days a week for 3 months. There were no significant changes in the fraction of cells with S-phase DNA content among the diploid cells. Cells harvested after 1 month of treatment with PG had a 50% reduction in the DNA production in tetraploid cells. DNA production was reduced to 20% after 2 months of treatment and was 0% at 3 months. The number of tetraploid cells was nearly 0 at the end of the 3-month treatment period. PG caused a disturbance in the proliferative response of rat urinary bladder epithelial cells. Mochida and Gomyoda (1987) investigated the effects of doses of PG on cul tured human KB cells. According to methods described by Mochida et al. (1983), the 50% inhibitory dose of PG was 0.31 M in cells incubated with the compound for 72 h.

Immune System Effects Denning and Webster (1987), studied the in vitro effects ofPG on human natural killer cell and neutrophil cell populations. In the first half of the experiment, a natural killer cytotoxicity assay was performed. Peripheral mononuclear cells were isolated from the blood of one human volunteer and used as the effector cells in the assay. Cultured K562 erythroleukemia cells were used as the target cells. Three concentrations of effector cells were incubated in separate wells of micro titer plates with PG in phosphate-buffered saline (PBS) diluted to final concen trations of 0.01, 0.1, and 1%. PBS alone was used as a control. One hundred microliters of the target cells was added to each well. Cytotoxicity was measured by observing the percentage 51 Cr release. The cytotoxicity of human natural killer cells was decreased significantly when cells were incubated with 1% PG (p < 0.002). Concentrations of 0.01 and 0.1% PG did not significantly decrease the natural killer cell activity. In the second part of the experiment, the same authors tested the effects of PG on isolated human neutrophils. Blood was collected from a human volunteer, and mononuclear cells were isolated. The isolated cells were incubated in a solution containing Hanks' Balanced Salt Solution, heparin, and 10 ~J.lluminol (5-amino- 2,3-dihydro-1,4-phthalazimedione). The incubated cells were taken in aliquots of 1 ml and placed in specially designed cuvettes for 30 min. The cells were then incubated with PG diluted to final concentrations ofO.O, 0.1, 0.5, or 1.0%. Latex particles coated with IgG were added to each tube, and the chemiluminescence was measured. The neutrophil chemiluminescence was significantly decreased in the cells incubated with 0.5 and 1% PG.

Other Biological Effects Floersheim (1985) investigated the effects of PG on the metabolism of ethanol in mice. Four groups offour female albino MAG mice were injected with 0.0, 0.05

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

448 COSMETIC INGREDIENT REVIEW ml/10 g, 0.025 ml/10 g, or 0.01 ml/10 g PG 30 min prior to the intraperitoneal administration of 4,500 mg/kg body wt ethanol. The survival rate for the mice given 0.01 ml PG was 19%. The survival rates in the mice fed the two larger doses were 75% for the mice given 0.05 ml PG and 87% for the mice given 0.025 ml PG. The survival rates in the mice given the two larger doses were significantly greater (p < 0.001) than in the mice given the lowest dose. The effects of PG on the metabolism of enflurane anesthesia were investigated by Fish et al. (1988). Thirty-two male Fischer strain rats were divided into four groups. Saline was administered to 10 control rats, 10 rats received 0.4 mg/ml etomidate, 6 received 7% (vol/vol) PG intravenously, and 6 received PG plus etomidate. All rats were then exposed to 2% enflurane vapor in an anesthesia chamber for a period of 1 h. Serum samples were collected at 0, 2, 4, and 24 h after exposure to the vapor. The metabolism of the anesthesia was measured by anal ysis of rat serum F- . There was a significant decrease in the rate of metabolism of enflurane in rats pretreated with PG as compared with saline controls (p < 0.05). In a second experiment, the same authors tested the effects of PG administra tion on adrenal steroidogenesis in the rat. Four groups of male Fischer strain rats, five per group, were given intravenous injections of saline (controls), 0.4 mg/ml etomidate, 7% (vol/vol) PG, or etomidate plus PG. Thirty minutes later, 0.2 JJ.g adrenocorticotropic hormone (ACTH) was administered to all groups via intra venous injection, causing aldosterone to be released into the serum. Serum sam ples were drawn at 30, 60, and 90 min post-ACTH administration. Administration of PG did not cause any significant changes in adrenal steroidogenesis in the rat in this experiment. The effects of intraperitoneal doses of PG on the metabolism of Hexobarbital and Zoxazolamine in the rat were studied by Dean and Stock (1974). Two groups of rats, three per group, were given doses of 4 ml/kg PG twice a day for 3 consecutive days. Two groups of three rats each served as untreated controls. Each group was then given either 125 mg/kg Hexobarbital (to test effect on sleep ing time) or 120 mg/kg Zoxazolamine (to test effect on paralysis time). The rats pretreated with PG and given Hexobarbital had significantly greater sleeping times than the control rats (p < 0.01). The treated rats given Zoxazolamine had significantly longer paralysis times than the control rats (p < 0.01). The effects of PG on antipyrine metabolism in humans were investigated by Nelson et al. (1987). Ten healthy volunteers were given 5-ml doses ofPG orally 10 min prior to administration of 1.2 g antipyrine. Additional doses of PG were administered every 4 h for 44 h (total PG dose = 55 ml). Plasma samples were taken from the venous cannula at 0, 5, 10, 15, 30, and 45 min and 1, 2, 4, 6, 8, 10, 12, 24, and 32 h after PG administration. Plasma antipyrine concentrations were measured by high-performance liquid chromatography. The authors found no significant effects on antipyrine clearance in humans due to PG administration at the doses investigated in this experiment. Male Sprague-Dawley rats (number not stated) were given 4 mllkg/day PG via gastric intubation for 30 days in a short-term study by Hoenig and Werner (1980). A second group of animals was given distilled water and served as the control.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 449

Body weights were measured throughout the experiment. On day 31, blood sam ples were taken from the carotid artery, and the livers of the animals were re moved for lipid analysis. There were no significant differences in body weights or liver weights in the PG-treated animals as compared with controls. There was a significant increase in total cholesterol in the livers of the PG-treated animals (p = 0.02). The effect of PG on liver structure was evaluated using 15 male Wistar rats (5 weeks old). PG [15% (voVvol) in drinking water] was administered to the animals daily by gastric intubation, and animals were killed at time periods ranging from 1 week to 3 months after initiation of the study. Thin sections of the liver, fixed with glutaraldehyde, were examined using an electron microscope. PG adminis tration resulted in remarkably enlarged mitochondria with well-developed cristae and a dense matrix in every hepatocyte. These changes were noted throughout the hepatic lobule, but were more distinct in hepatocytes of the peripheral zone of the hepatic lobule. Other ultrastructural changes in hepatocytes included proliferation of smooth-surfaced endoplasmic reticulum and an increase in the number of ly sosomes and microbodies (Wakabayashi et al., 1991). Enzyme effects of short-term administration of PG in rats were investigated by Amma et al. (1984). Two groups of six male and six female rats were given oral doses of 1 rnl28.4% PG/100 g body wt daily for 30 days. A control group of 12 rats was given daily doses of distilled water. On day 31, blood samples for enzyme studies were collected from the orbital sinus of each animal. Glucose-6- phosphate: NADP oxidoreductase, blood' catalase activity, and whole-blood re duced glutathione (GSH) were measured. The growth rate of test animals was not significantly different from controls. In the treated animals, the whole-blood GSH was significantly decreased (p .::; 0.001), and erythrocyte catalase activity was significantly decreased in the males (p .::; 0.01). Ahluwalia and Amma (1984) studied the effects of short-term administration of PG on rat erythrocyte morphology and function. One group of six male rats was given oral doses of 284 ILl PG daily for 30 days. A second group of six rats serving as a control group was given oral doses of distilled water. On day 31, blood samples were drawn from the orbital sinus of each rat, and membrane lipids were extracted from erythrocytes. Short-term oral doses of PG caused a significant decrease in total lipids (p < 0.001), cholesterol (p < 0.05), and phospholipids (p < 0.05) in the rat erythrocytes. The decrease in lipids altered the erythrocyte mor phology in the PG-treated animals. Christopher et al. (1989) investigated the effects of oral PG administration on the development of Heinz body hemolytic anemia in cats. Two groups of cats were fed commercial diets in the conditioning phase of the experiment. Control values for Heinz bodies were obtained. Each cat served as its own control. Six of the cats were fed diets containing 12% (1.6 g/kg/day) PG for 5 weeks (low-dose study). Five adult cats were fed diets containing 41% (8 g/kg/day) PG for 22 days (high-dose study). Blood samples were collected and Heinz body formation was investigated. The percentage of Heinz bodies present during the control period was 2.98 ± 0.90% in the low-dose group. Mter 1 week on the 12% PG diet, the percentage of Heinz bodies was significantly increased to 11.33 ± 3.36% (p <

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

450 COSMETIC INGREDIENT REVIEW

0.002). After 35 days, the percentage of Heinz bodies was 28%. In the high-dose group, the control value for percentage of Heinz bodies was 4.8 ± 1.81%. The percentage of Heinz bodies was 92.1 ± 3.6% on day 9. This value was significantly increased compared with control values (p < 0.001). Singh et al. (1982) investigated various pharmacologic parameters of PG in a series of animal studies. Swiss mice (number not stated) in a locomotor activity experiment were given either intraperitoneal or oral doses of 10, 20, 50, or 100% PG. Locomotor activity was measured at 15, 30, 60, and 120 min postdosing. In animals given 100% PG intraperitoneally, there was a significant reduction in locomotor activity. The effect of PG administration on forced locomotor activity was also studied in Wistar rats and Swiss mice (number not stated). Mice given peroral doses of 100% PG had a 60% reduction in activity compared with controls. Upon intra peritoneal administration of the compound, there was a significant decrease in forced locomotor activity in both rats and mice given 100% PG. With use of the same doses in an inclined plane test involving mice, a smooth wooden plate at an angle of 45° was used to determine ataxia. By both routes of administration, 20-100% PG caused ataxia in mice. Further investigations included the effects of PG on rat and guinea pig smooth muscle samples. Pieces of the ileum from a guinea pig were placed in an organ bath and 0.5 ml of 10, 20, 50, or 100% PG was added for 3 min. Contractions were stimulated in the muscles by addition of acetylcholine chloride, histamine dihy drochloride, or barium chloride. At concentrations above 20%, a nonspecific, dose-dependent inhibition of contraction was noted. In the rat, uterine samples were immersed in organ baths and PG (same concentrations) were added. Con tractions were induced in the smooth muscle samples by addition of 5-hydroxy tryptamine. All PG concentrations produced a nonspecific dose-dependent inhi bition of contractions. PG did not produce any paralytic effects, analgesic activ ity, or hypnotic effects in mice or rats at any of the concentrations tested. Several other investigators also studied the effects of PG on muscle function. The potential for 40% (vol/vol) PG in distilled water to cause skeletal muscle damage after intramuscular injection was evaluated in vitro using isolated exten sor digitorum longus muscles from rats. The muscles were injected with 15 I-Ll of the solution and placed in a balanced salt solution bubbled with 95% 0 2/5% C02 • The myotoxicity of injected solutions was evaluated by noting the cumulative release of creatine kinase (intracellular cytosolic enzyme) into the incubation medium over a 2-h period. Creatine kinase levels and histological evaluation are the commonly used indexes of skeletal muscle damage, both clinically and exper imentally. Cumulative creatine kinase release increased as a function of time throughout the 2-h incubation period (Brazeau and Fung, 1989). The results of this study were then evaluated with in vivo studies on creatine kinase activity in five male New Zealand white rabbits (weights 2.0-4.2 kg). The animals received a 1-ml injection of each of the following: 40% (vol/vol) PG, 40% (vol/vol) polyethylene glycol 400, and normal saline. The solutions were injected into the mid-lumbar muscles of each animal in a three-way crossover random design; there was a minimum of 2 weeks between each phase of the experiment.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 451

The injection sites randomly rotated such that no single muscle area received more than one injection, and blood samples were obtained from the central artery or marginal ear vein up to 72 h postinjection. The corrected area under the serum kinase activity versus time curve following propylene glycol injection was statis tically larger (p < 0.05) than those after the injection of polyethylene glycol and normal saline (Brazeau and Fung, 1989). Kaldor et al. (1971) studied the effects of PG and various PPGs on the contrac tion of rabbit muscles. Glycerol-extracted fibers of rabbit psoas muscle were cut into 50-mm bundles. Half of each bundle was incubated with 0.2 M KCl/0.05 M Tris (the incubation mixture), and 6.6 x 10-5 M Mg-ATP was added to initiate a reaction. The second half of the fiber bundle was incubated with the incubation mixture and the following test substances: 16 mg/ml PG, 33 mg/ml PG, 66 mg/ml PG, 33 mg/ml PPG 225, 66 mg/ml PPG 225, 33 mg/ml PPG 425, or 66 mg/ml PPG 425. All doses of PG, PPG 225, and PPG 425 promoted shortening of the test muscle fibers, as well as development of tension within the test fibers. An aqueous solution of PPG 400 (dose 20 mg/kg) was administered intrave nously to a dog (weight not stated) that was curarized, but not anesthetized. The results of an electroencephalogram (EEG) indicated increased electrical activity in all lobes of the brain. The increased activity was first noted in the frontoparietal region and spread gradually to temporal and occipital regions. Following intrave nous administration of a PPG 750 aqueous solution and an aqueous solution of PPG 1200 (doses 10 mg/kg), the same EEG pattern was noted. However, the spread of electrical activity to other areas of the brain was more rapid. Electro cardiogram patterns following the administration of PPG 400 and PPG 750 indi cated positive chronotropic responses. However, such responses were not ob served after administration of PPG 1200 or PPG 2000 (Shideman and Procita, 1951). The administration of aqueous PPG 400, PPG 750, or PPG 1200 to the anesthe tized, noncurarized dog (weight not stated) at doses of 5 mg/kg caused an increase in respiratory rate and amplitude. Manifestations of increased central nervous system activity in the form of enhanced stretch reflexes, muscle tremors, and movements were also noted. At a higher dose, 25 mg/kg, convulsant activity was noted (Shideman and Procita, 1951). The intravenous administration of 10- and 50-mg/kg doses of PPG 425, diluted 1: 10 in saline solution, to groups of eight anesthetized, adult mongrel dogs (weights not stated) caused acceleration of cardiac and respiratory rates. Effects of the smaller dose on the electrocardiogram pattern ranged from a barely per ceptible change to the presence of occasional ventricular extrasystoles. Intrave nous doses of 50 mg/kg resulted in violent clonic convulsions and increased the heart rate to nearly double that of the control value; respiration was described as shallow and spasmodic. Doses of PPG 1025 (1 g/kg) and PPG 2025 (5 g/kg) pro duced no appreciable effect on the electrocardiogram within 15 min postdosing (Shaffer et al., 1951). The effects of PG on cardiopulmonary function were studied by a number of investigators. An intravenous bolus injection (0.5 ml/kg) of 30% PG in sterile water caused a transient significant increase in pulmonary arterial pressure in a

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

452 COSMETIC INGREDIENT REVIEW

group of eight mixed breed sheep (weights 20--28 kg). The peak mean pulmonary arterial pressure increased from 17 ± 1 to a peak of35 ± 4 mm Hg (p < 0.05), and significant increases in pressure were not noted at 5 min postdosing. A similar finding was reported in lambs, but not in dogs or guinea pigs (Quinn et al., 1990). Eichbaum and Yasaka (1976) studied the antiarrhythmic effects of solvents, including PG. Twenty rats and 15 dogs were anesthetized prior to having arrhyth mias (monitored by electrocardiograph) induced via ouabain injections. The ani mals were then dosed intravenously with 0.2 ml/kg body wt 100% PG or 0.2 mVkg body wt 70% PG. At each dose administered, PG produced either a temporary or a permanent change to a normal cardiac rhythm. These authors also investigated the local and general anesthetic effects of PG in rats, dogs, and rabbits. Four rats were given periarticular injections of0.1 ml 70% PG. No significant local anesthetic effect was noted in any of the test animals. Four rats, two rabbits, and two dogs were given intravenous injections of up to 4 ml/kg body wt 100% PG (specific doses not stated). There were no significant general anesthetic effects noted in any of the animals within 60 min after dosing. Thomas et al. (1980) investigated the antibacterial properties ofPG. Agar plates were inoculated with the test organisms Staphylococcus aureus, Pseudomonas aeruginosa, Streptococcus faecalis, and Streptococcus pyogenes. After a 24-h incubation period, a 1.5-cm hole was cut in the center of each agar plate. A 100% solution of PG was poured into the hole, and the zones of inhibition of the test organism's growth were measured on each plate. PG did not produce any zones of inhibition in any of the plates. PG did not have antiseptic properties in this test.

TOXICOLOGY Acute Inhalation Toxicity Five male albino rats (weights 213-238 g) were exposed for 7 h to vapor satu rated with PPG 1200. Gross pathologic examinations were conducted on one animal at day 3 postexposure and on another animal at the end of the study. There were no signs of toxicity in animals dosed with vapor generated at room temper ature. Two of the animals exposed to vapors generated at 100°C had lung rales on day 10 postexposure and one rat had a bloody discharge on day 17. Weight gain was noted in all animals. Emphysema and a moderate degree of hepatic and renal involvement were noted in the animal killed on day 3 and in the animal killed at the end of the study, respectively (FDA, 1992).

Acute Oral Toxicity Propylene Glycol

Sax (1979) reported an acute oral LD50 of 21 g/kg body wt for PG in rats. Five female Fischer strain rats were administered doses of PG to determine the oral LD50 of the compound. The 24-h LD50 ofPG was 22.8 g/kg body wt. The 95% confidence interval was 20.5-25.3 g/kg. The lowest recorded 24-h lethal dose in this experiment was 20.9 g/kg, and the lowest recorded 48-h lethal dose was 19.8 g/kg (Clark et al., 1979).

JAm Coli Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 453

Bartsch et al. (1976) reported an oral LD50 of 25 ml/kg in 10 male and female Sprague--Dawley rats. The 95% confidence level was 19.2-32.5 ml/kg. Saini et al. (1987) investigated the effects of acute oral administration of PG using female Wistar strain rats. Three groups of rats were given 1-ml oral doses of 9.66, 19.32, or 38.64 mmol/kg body wt PG. Blaod samples were drawn from the optical sinus at 0 h, 30 min, 60 min, 120 min, and 24 h postdosing. Acute oral administration of PG caused significant decreases in the levels of fibrinogen, albumin, and globulin in the plasma. PG affected the function of the liver in either the synthesis or the secretion of proteins in the rat. The accidental oral administration of 3.8 L of PG to a male horse (weight 450--500 kg) resulted in pain, ataxia, salivation, and excessive sweating within 10--15 min. With the exception of ataxia, all clinical signs had resolved within 5 min, and an appropriate treatment protocol was begun. On the next day, the animal became increasingly ataxic and died of apparent respiratory arrest. Death occurred at 28 h postingestion. No appreciable gross lesions were noted at nec ropsy; PG concentrations of 9,000 mg/L in serum and 7,500 mg/L in renal fluids (combined urine and blood) were detected. Microscopic changes included mod erate myocardial perivascular edema with dilatation of lymphatics and moderate pulmonary edema characterized by proteinaceous material in alveoli and in some of the bronchioles. Hepatic lesions consisted of scattered single-cell hepatocytic necrosis and minimal acute suppurative pericholangitis. Peracute renal infarcts characterized by multiple linear areas of coagulative tubular necrosis were also observed (Dorman and Haschek, 1991).

Polypropylene Glycols

The acute oral toxicity of PPGs (20% aqueous) was evaluated using three groups offive male albino rats of the Sherman strain (weights between 90 and 120 g). The mean LD50 values reported for the three groups were as follows: PPG 425 (2.91 g/kg), PPG 1025 (2.15 g/kg), and PPG 2025 (9.76 g/kg). Sluggishness, pros tration, tremors, convulsions, and rapid death were noted in all treatment groups. The following gross lesions were observed in animals that died: minor pulmonary hemorrhage, congestion of the liver and spleen, renal ischemia, and injection of intestinal blood vessels (Shaffer et al., 1951). In another study, 121 male rats (seven groups, 13-20/group) received various oral doses of 10% aqueous PPG 1200. The LD50 was 640 mg/kg (confidence limits 376--1,088 mg/kg) (FDA, 1992). Additional acute oral toxicity studies on PPGs of various molecular weights (300--3,900) have indicated LD50 values (rats) ranging from 0.5 to >40 g/kg (Amer ican Industrial Hygiene Association, 1980). An aqueous solution of PPG 1200 was tested in an acute oral toxicity study involving 10 dogs. Neither the test substance concentration nor the weight range of animals tested was stated. A subconvulsant effect was noted after the admin istration of 50 mg/kg (Shideman and Procita, 1951). The acute oral toxicity of 50% aqueous PPG 1200 was evaluated using 55 male and 63 female guinea pigs (six groups/sex; 7-15/group). The LD50 values for male

JAm Coil Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

454 COSMETIC INGREDIENT REVIEW and female guinea pigs were 1,320 mg!kg (95% confidence limits 930--1,875 mg/kg) and 1,420 mg/kg (95% confidence limits 1,090--1,850 mg/kg), respectively (FDA, 1992). Acute oral toxicity studies on PPGs of various molecular weights have indicated LD50 values (guinea pigs) ranging from 1.5 to 17 g/kg (American Industrial Hy giene Association, 1980).

Acute Dermal Toxicity

The dermal toxicity of PPGs (undiluted) was evaluated according to the proce dure ofDraize et al. (1944) using groups of five rabbits. Doses ofPPG 425 (10 and 20 ml/kg), PPG 1025 (20 ml/kg), and PPG 2025 (20 mllkg) were applied for 24 h, and animals were observed for 13 days. All animals dosed with PPG 1025 and PPG 2025 survived the 24-h contact period and the observation period. Two of the five rabbits dosed with PPG 425 (20 mllkg) and one of five rabbits dosed with 10 mllkg PPG 425 died; animal deaths resulted from dermal penetration of the test sub stance. It was concluded that PPG 1025 and PPG 2025 did not penetrate the skin readily and that PPG 425 penetrated the skin to some extent (Shaffer et al., 1951). In another study, the dermal toxicity of undiluted PPG 1200 was evaluated using 10 albino rabbits (weights 2.43-3.82 kg). The distribution of doses applied was as follows: 1.0 ml/kg body wt (two rabbits), 3.0 mllkg (two rabbits), 10.0 mllkg (one rabbit), and 30.0 mllkg (five rabbits). Doses (24-h exposures) were applied under a heavy-gauge plastic cuff held in place with rubber bands and a cloth bandage that was taped securely to the marginal hair. At the end of exposure, sites were washed with soap and water, rinsed thoroughly, and dried with a cloth towel. The animals were observed for signs of toxicity during exposure and periodically for 2 weeks after exposure. Doses of 1.0, 3.0, or 10.0 ml/kg did not cause death, and, at most, moderate edema and slight erythema were observed at application sites. Body weight gains were noted in all animals, including the 30-ml/kg dose group. One of five rabbits in the 30.0-ml/kg group died, and the remaining rabbits ap peared ill during the 2-week observation period. Overt signs of toxicity were not noted in these rabbits at the conclusion of the study (FDA, 1992). Following single 24-h dermal applications of PPGs to animals (dose 30 mllkg; species not stated), four of five animals survived treatment with PPG 400 or PPG 1200, and six of six animals survived treatment with PPG 200. All test substances were said to have been poorly absorbed through the skin (American Industrial Hygiene Association, 1980).

Acute Intraperitoneal Toxicity Propylene Glycol

In the rat, the reported intraperitoneal LD50 for PG was 13 g/kg (Sax, 1979). Bartsch et al. (1976) reported an intraperitoneal LD50 of 13.0 g/kg for PG in male and female Sprague-Dawley rats. The 95% confidence level was 10.2-16.5 ml!kg. Six male NMRI mice were given intraperitoneal doses of PG diluted in water to

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 455

10 ml/kg body wt. The intraperitoneal LD50 was 11.2 g/kg body wt. Confidence levels were 11.0-11.5 g/kg (Budden et al., 1979). Groups of five male and five female C3H strain mice were given intraperitoneal injections of 5 ml/kg PG. Mice were observed for 7 days. All had signs of lack of coordination followed by deep narcosis. There were no deaths during the 7-day observation period. No changes were noted in weight gain, feed and water con sumption, or general behavior. At necropsy, peritonitis was observed in the test animals (Hickman, 1965).

Polypropylene Glycols The acute intraperitoneal toxicity ofPPGs (undiluted) was evaluated using three groups of five male albino rats of the Sherman strain (weights between 90 and 120 g). The mean LD50 values reported for the three groups were as follows: PPG 425 (0.46 g/kg), PPG 1025 (0.23 g/kg), and PPG 2025 (4.47 g/kg). In all treatment groups, death was preceded by tremors, prostration, frothing at the mouth, and audible rales (Shaffer et al., 1951). Acute intraperitoneal LD50 values that have been reported for aqueous PPG solutions are as follows: PPG 400 (700 ± 75 mg/kg, 50 mice), PPG 750 (195 ± 8 mg/kg, 50 mice), PPG 1200 (113 ± 12 mg/kg, 60 mice), and PPG 2000 (3,600 ± 400 mg/kg). Weight ranges (not stated) were similar between all treatment groups (Shideman and Procita, 1951). The results of other acute intraperitoneal toxicity studies have indicated LD50 values (rats) of 0.1-0.7 g/kg for PPGs ranging in molecular weight from 400 to 1,200 and an LD50 of 3.6 g/kg for PPG 2000 (American Industrial Hygiene Asso ciation, 1980).

Acute Intravenous Toxicity Propylene Glycol

Bartsch et al. (1976) reported an intravenous LD50 for PG of 6.2 ml/kg in male and female Sprague-Dawley rats. The 95% confidence level was 5.2-7.4 ml/kg. The same authors reported an intravenous LD50 of 6.4 ml/kg in male and female SPF-NMRI strain mice. The 95% confidence level was 6.1-6.9 ml/kg.

Polypropylene Glycols The acute intravenous toxicity of PPGs (undiluted) was evaluated using three groups of five albino rats of the Sherman strain (weights between 90 and 120 g). The mean LD50 values reported for the three groups were as follows: PPG 425 (0.41 g/kg), PPG 1025 (0.12 g/kg), and PPG 2025 (0.71 g/kg). In all treatment groups, death was preceded by tremors, prostration, frothing at the mouth, and audible rales. Convulsions were noted only in groups treated with PPG 1025 and PPG 2025 (Shaffer et al., 1951). Aqueous PPG solutions were tested in acute intravenous toxicity studies in volving dogs (10 dogs/study; weights not stated). The results were as follows (Shideman and Procita, 1951): PPG 400 (dose 10-20 mg/kg, tremors with convul-

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

456 COSMETIC INGREDIENT REVIEW sions); PPG 750 (5-7 mg/kg, tremors), PPG 750 (8-15 mg/kg, dogs with convul sions survived); PPG 750 (20 mg/kg, dogs with convulsions died), PPG 1200 (5-7 mg/kg, tremors), PPG 1200 (15 mg/kg, dogs with convulsions survived), PPG 1200 (20 mg/kg, dogs with convulsions died), and PPG 2000 (100 mg/kg, no visible effects).

Acute Intramuscular Toxicity The acute intramuscular toxicity of aqueous solutions of PPG 1200 was evalu ated using two groups of 10 dogs (weights not stated). Subconvulsant effects were noted in dogs injected with 45 mg/kg PPG 1200, and mild convulsions were noted in dogs injected with 50--60 mg/kg PPG 1200 (Shideman and Procita, 1951). The results of acute toxicity studies on PG and PPG are summarized in Tables 7 and 8, respectively.

Short-Term Oral Toxicity The effects of short-term administration of PG in rats and dogs were studied by Staples et al. (1967). Charles River rats (10/group) were given PG via stomach tube three times a day for 3 days. Doses were 0.75 ml!kg 100% PG, 1.50 mllkg 100% PG, 3.0 mllkg 100% PG, 3.0 mllkg 75% PG, and 3.0 ml!kg 50% PG. Three adult mongrel dogs were also given PG via stomach tube for the same time period. Doses were 0.75, 1.5, or 3.0 ml/kg PG. One rat given the smallest dose and two rats given 3 mllkg 100% PG had slight hyperemia of the GI tract. One rat given 1.5 ml/kg 100% PG had severe hyperemia of the GI tract. No other compound-related effects were noted in the rats. The dog given 1.5 ml/kg PG had evidence of patches of slight hyperemia. No other adverse effects were noted in any of the dogs. The oral toxicity of PG was evaluated using a total of 11 male and female mongrel cats (weight range 3-4 kg) (Christopher et al., 1990b). Five cats were fed a diet containing 12% PG (dry weight) for 5 weeks; this diet was preceded by 4 weeks on a control diet containing 12% PG (dry weight). The remaining six cats were fed a diet containing 41% PG (dry weight) for 22 days. Mean values for

TABLE 7. Reported LD50 values for Propylene Glycol

Animal Strain Sex Dose Results Ref. Rat Oral 21 g/kg Sax (1979) Rat Fischer F Oral 22.8 g/kg Clark et al. (1979) Rat Sprague-Dawley M,F Oral 25 ml/kg Bartsch (1976) Rat Oral 27 g/kg Layton et al. (1987) Rat i.p. 13 g/kg Sax (1979) Rat Sprague-Dawley M,F i.p. 13 mllkg Bartsch (1976) Rat SPF M,F i.p. 13.7 mllkg Hickman (1965) Mouse NMRI M i.p. 11.2 g/kg Budden et al. (1979) Mouse SPF-NMRI M,F i.p. 9.3 mllkg Bartsch (1976) Rat Sprague-Dawley M,F i.v. 6.2 mllkg Bartsch (1976) Mouse SPF-NMRI M,F i.v. 6.4 mllkg Bartsch (1976) Mouse i.v. 8.0 g/kg Sax (1979) Rat i.m. 20.0 g/kg Sax (1979) Mouse S.C. 18.5 g/kg Sax (1979)

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 457

TABLE 8. Acute toxicity studies on Polypropylene Glycol (PPG)

Animal Strain Sex Dose Results Ref.

Rat Sherman M Oral PPG 425: LD50 = 2.91 g/kg Shaffer et al. (1951) Rat Sherman M Oral PPG 1025: LD50 = 2.15 g/kg Shaffer et al. (1951) Rat M Oral PPG 1200: LD50 = 640 mg/kg FDA (1992) Rat Sherman M Oral PPG 2025: LD50 = 9.76 g/kg Shaffer et al. (1951) Dog Oral Subconvulsant effect at 50 Shideman and Procita mg/kg (1951) Guinea pig M,F Oral PPG 1200: LD50 = 1,320 American Industrial mg/kg (males) and 1,420 Hygiene Association mg/kg (females) (1980) Rabbit Dermal PPG 425: 20% and 40% Shaffer et al. (1951) mortality at 10 and 20 ml/kg Rabbit Dermal PPG 1025: no deaths at 20 Shaffer et al. (1951) ml/kg Rabbit Dermal PPG 1200: 20% mortality at FDA (1992) 30 ml/kg Rabbit Dermal PPG 2025: no deaths at 20 Shaffer et al. (1951) mlJkg Mouse i.p. PPG 400: LD50 = 700 mg/kg Shideman and Procita (1951) Mouse i.p. PPG 750: LD50 = 195 mg/kg Shideman and Procita (1951) Mouse i.p. PPG 1200: LD50 = 113 mg/kg Shideman and Procita (1951) Mouse i.p. PPG 2000: LD50 = 3,600 Shideman and Procita mg/kg (1951) Rat Sherman M i.p. PPG 425: LD50 = 0.46 g/kg Shaffer et al. (1951) Rat Sherman M i.p. PPG 1025: LD50 = 0.23 g/kg Shaffer et al. (1951) Rat Sherman M i.p. PPG 2025: LD50 = 4.47 g/kg Shaffer et al. (1951) Rat Sherman i.v. PPG 425: LD50 = 0.41 g/kg Shaffer et al. (1951) Rat Sherman i.v. PPG 1025: LD50 = 0.12 g/kg Shaffer et al. (1951) Rat Sherman i.v. PPG 2025: LD50 = 0.71 g/kg Shaffer et al. (1951) Dog i.v. PPG 400: tremors with Shideman and Procita convulsions, at 10-20 (1951) mg/kg Dog i.v. PPG 750: tremors at 5-7 Shideman and Procita mg/kg (1951) Dog i.v. PPG 750: convulsions but no Shideman and Procita deaths at 8-15 mg/kg (1951) Dog i.v. PPG 750: convulsions and Shideman and Procita deaths at 20 mg/kg (1951) Dog i.v. PPG 1200: tremors at 5-7 Shideman and Procita mg/kg (1951) Dog i.v. PPG 1200: convulsions but no Shideman and Procita deaths at 15 mg/kg (1951) Dog i.v. PPG 1200: convulsions and Shideman and Procita deaths at 20 mg/kg (1951) Dog i.v. PPG 2000: no visible effects Shideman and Procita (1951) Dog i.m. PPG 1200: subconvulsant Shideman and Procita effects at 45 mg/kg (1951) Dog i.m. PPG 1200: mild convulsions Shideman and Procita at 50-60 mg/kg (1951)

JAm Coli Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

458 COSMETIC INGREDIENT REVIEW

ingested PG were 1.6 g/kg body wt/day for the 12% dose group and 8.0 g/kg body wt/day for the 41% dose group. In the low-dose group, plasma o-lactate (plasma samples obtained only from two cats) increased from 0.08 ± 0.03 mmol/L (range 0-0.16 mmol/L) during the control period to 1.96 ± 0.75 mmol/L (range 0.2-4.2 mmol/L) on day 24. In the high-dose group, plasma o-lactate increased rapidly to 4.21 ± 1.96 mmol/L by day 10 (p < 0.001), and on day 24 a mean value of7.12 ± 0.14 mmol/L was reported. These findings indicate that PG is metabolized in vivo, at least in part, to o-lactate. The results of physical examinations performed on cats that received the smaller dose did not indicate any abnormal findings. Ob servations in cats that received the high dose included moderate polyuria and polydypsia; these findings are consistent with renal excretion of PG, which acts as an osmotic diuretic (Hanzlik et al., 1939). Cats in the high-dose group also devel oped decreased activity, mental depression, and slight to moderate ataxia. These observations may be related to the metabolism of PG to o-lactate. For example, encephalopathy in humans, characterized by disorientation and ataxia, is associ ated with o-lactate concentrations of >4.0 mmol/L (Ludvigsen et al., 1983).

Short-Term Intravenous Toxicity Fort et al. (1984) investigated the effects of short-term administration of PG using rats and dogs. Groups of five male and five female rats were given intrave nous infusions ofPG/ethanol/water (5:1:4) over a period of2 weeks. The dose was 5 ml/kg/day at a rate of 0.3 ml/min. One male beagle and one female beagle dog were given the same PG/ethanol/water mixture at doses of 4 ml/kg/day at a rate of 6 ml/min. Excretion of red urine was noted after the initial dose in both test species and continued throughout the 2-week study in most rats and in some dogs (number not specified). Mter the 2-week treatment period, decreases in packed cell volume, hemoglobin, and erythrocyte counts were noted in both test species. At necropsy, splenomegaly and renal hemosiderin deposits were observed in the rats.

Subchronic Oral Toxicity Propylene Glycol Gaunt et al. (1972) investigated the effects of subchronic administration of PG using rats. Fifteen male and 15 female Charles River CD rats were fed diets containing 0 (control) or 50,000 ppm PG for 15 weeks. The PG administered in the feed corresponded to a dose of 2.5 g/kg/day in the test animals. Blood samples were collected at the end of the study, and a hematological examination was performed. The brain, heart, liver, spleen, kidneys, adrenal glands, gonads, and pituitary gland were collected from each animal, and organ weights were com pared with those of controls. No significant differences were found between the control and PG-dosed animals. No compound-related lesions were noted.

Polypropylene Glycols PPG 2000 was administered to rats (weights, strain, and number tested not stated) over a period of 100 days. Concentrations of0.1, 0.3, 1.0, and 3.0% were

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 459 administered in oral doses of ~5()....1,500 mg/kg/day. Adverse effects were not noted at concentrations of 0.1--1.0%. Slight decreases in growth were observed after the administration of 3% PPG 2000 (American Industrial Hygiene Associa tion, 1980). In a 90-day study, PPG 2000 was administered orally to rats (weights and number tested not stated) in doses ranging from 275 to 501 mg/kg/day. There was no evidence of adverse histopathologic, hematologic, or clinical chemistry effects in any of the animals tested. Body weight effects (not specified) were noted at the highest dose tested. Similar results were reported for dogs (weights and number tested not stated) that received doses of PPG 2000 ranging from 526 to 810 mg/ kg/day (American Industrial Hygiene Association, 1980). In another study, PPG 1200 was fed to three groups of four purebred beagle dogs (weights 5.95-12.5 kg) for 90 days. Each group consisted of two male and two female dogs. The three groups ingested PPG 1200 at concentrations of 0.1, 0.3, and 1.0% in the diet, respectively; six beagles in the control group were fed diets without PPG. At the highest "no effect level" (0.3%), average daily con sumption doses were 79 and 99 mg/kg for two male dogs and 90 and 123 mg/kg for two female dogs. Data based on the following analyses did not indicate adverse effects at any of the concentrations tested: demeanor, feed consumption, hema tologic parameters, clinical chemistry determinations, urinalysis, organ weights, organ/body weight and organ/brain weight ratios, or gross and microscopic ex amination of tissues and organs. Similar results were reported when three groups of 50 rats were dosed with 0.1, 0.3, and 1.0% PPG 1200, respectively, according to the same procedure. Each group was equally mixed and weight ranges for male and female rats were 321-334 and 218-233 g, respectively (FDA, 1992). PPG 750 was administered to rats (weights and number tested not stated) over a period of 100 days. Concentrations of 0.1 and 1% were administered at doses of ~50 and 500 mg/kg/day. PPG 750 (0.1%) did not induce any adverse effects. However, in the group dosed with 1% PPG 750, there was a slight increase in liver and kidney weights; there were no histological changes. Neither of the doses resulted in a central nervous system stimulatory effect (American Industrial Hy giene Association, 1980).

Subchronic Dermal Toxicity Propylene Glycol PG (0.2 ml) was injected subcutaneously into 12 male hairless mice of the hrlhr Oslo strain (1()....12 weeks old) three times per week for 3 months. None of the animals died and there was no ulceration at the injection sites. Based on micro flow histograms of urinary bladder epithelial cells (Farsund, 1974), the following results were reported: The proportion of diploid cells in the S-phase fluctuated but was not significantly altered. The proportion of tetraploid cells in the S-phase was significantly reduced, and at 3 months, there was no DNA synthesis in these cells. Overall, the proportion of diploid cells increased, the number of tetraploids was slightly reduced, and almost all of the octoploid cells disappeared. The alterations reported were described as disturbed regenerative reactions resulting from PG-

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

460 COSMETIC INGREDIENT REVIEW induced acute cellular toxicity. Some of the bladder epithelial cells were killed and the mechanism of repeated DNA synthesis was disturbed (Farsund, 1978).

Polypropylene Glycol In a 3-month study, PPG 2000 was applied to the skin of rabbits (weights and number tested not stated) at doses of 1, 5, and 10 mllkg. Applications were made 5 days per week (24 h/day). Doses of 1 ml/kg did not cause any adverse effects. Slight depression of growth was noted after the administration of 5- and 10-ml/kg doses (American Industrial Hygiene Association, 1980).

Chronic Oral Toxicity In a chronic feeding study by Gaunt et al. (1972), PG was administered in the diet to rats for 2 years. Groups of 30 male and female Charles River CD rats were fed diets containing 0, 6,250, 12,500, 25,000, or 50,000 ppm PG. Feed and water were given ad libitum. General behavior was observed, and body weights were recorded every 2 weeks. At 104 weeks, animals were killed and brain, heart, liver, spleen, kidneys, adrenal glands, stomach, small intestines, and cecum were weighed. Samples of organs that appeared to have lesions were retained for mi croscopic examination. No significant compound-related lesions were noted in the test animals; all organ and body weights in the treated animals were not signifi cantly different from controls. Weil et al. (1971) investigated the effects of chronic administration of PG using dogs. Groups of five male and five female dogs were given diets containing 2 to 5 g/kg PG. Body weights were recorded throughout the experiment. At week 104, all animals were killed and organs were examined microscopically for lesions. At the 5-g/kg dose levels, dogs gained more weight than controls, and urine output was increased. No compound-related lesions were noted. The no-effect level for PG was 2 g/kg body wt. A group offour female dogs (weights 6-21 g) received 5% PG (in drinking water) for a period ranging from 5 to 9 months. The animals were allowed to drink the mixture twice daily during a period of 1 h. Additionally, each of four male dogs (weights not stated) was allowed to drink 600 ml of 10% PG (in drinking water) daily over a period ranging from 5 to 6 months. The three tests of liver function included galactose excretion, uric acid excretion, and the rose bengal content of blood. The excretion of phenolsulfonphthalein during a 2-h test period served as the test for renal function. The results indicated no impairment of hepatic or renal function in either male or female dogs (Van Winkle and Newman, 1941).

Ocular Irritation Propylene Glycol Clark et al. (1979) evaluated the ocular irritation potential of solar heat transfer fluids, including PG. PG (0.1 ml) was instilled into the conjunctival sac of one eye of each of six rabbits, and reactions were scored according to the methods of Draize et al. (1944). Eyes were examined at 1, 24, 48, 72, and 96 h posttreatment.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 461

At 24 h, a score of0.7 (maximum score = 110) was noted. At 72 h, the score was 0.3. At 96 h, the score was 0.0. There was no corneal damage noted in any of the test animals. The ocular irritation potential of 56 substances, including PG, was investigated by Guillot et al. (1982a). PG (pH 8.8; 0.1 ml) was instilled into the conjunctival sac of one eye of each of six male New Zealand White rabbits. Untreated eyes served as controls. Eyes were not rinsed after instillation. The acute ocular irritation index (AOI) (maximum possible score = 110) and the mean ocular irritation index (MOl)(maximum possible score = 40) were calculated for PG. The AOI was 11.33 and the MOl was 0.83. PG was a slight ocular irritant in this test. Douglas and Spilman (1983) conducted an in vitro ocular irritancy test using several substances, including PG. A 51 Cr release assay using a cultured corneal endothelial cell line was used to approximate ocular toxicity in vivo. Cells were preincubated with 0.5 f-LCi/ml 51 Cr for 1 h. The corneal endothelial cells were then incubated with the test substance. Different molar concentrations of the test sub stance were used to obtain the estimated dose that resulted in 50% toxicity (ED50). The test was run three times, and the ED50 values measured were 30, 13, and 13 M. The 95% confidence limits were 15-61, 7.2-24, and 8.2-20 M, respectively. PG was a nonirritant in this test.

Polypropylene Glycols

Following instillation of an excess ofPPG 425, PPG 1025, and PPG 2025 into the conjunctival sacs of rabbits, trace injuries to one or two of five eyes tested with each test substance were observed. PPG 425, PPG 1025, and PPG 2025 were classified as harmless agents (Shaffer et al., 1951). The ocular irritation potential of undiluted PPG 1200 was evaluated using an albino rabbit. The test substance (0.1 ml) was instilled into the conjunctival sac of both eyes; only one eye was rinsed after instillation. Both eyes were examined for conjunctival irritation, corneal injury, and internal effects such as iritis and len ticular damage. At most, the test substance induced very slight discomfort and slight transient conjunctival irritation. Signs of ocular irritation had cleared by 24 h postinstillation (FDA, 1992).

Skin Irritation/Sensitization Propylene Glycol

The skin irritancy potential of PG was studied by Lashmar et al. (1989) using nude mice. PG (10, 25, or 50%) was instilled in polyvinyl chloride cups (vol 0.3 cm2) on the dorsal side of three mice. The test substance remained in contact with the skin for 24 h. At the end of 24 h, the animals were killed and a sample of the exposed skin was taken for microscopic examination. Reactions were scored according to the methods of Ingram and Grasso (1975) (maximum possible score = 77). At 10%, no irritation was observed. At 25%, the score was 5, and at 50%, the score was 11. In the animals exposed to 50% PG, hypertrophy, dermal in-

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

462 COSMETIC INGREDIENT REVIEW

flammation, and proliferation stimuli were noted, indicating that this concentra tion of PG may cause skin irritation. In a mouse external ear sensitization assay, 100% PG was applied to the right ear of each of 19 mice (Swiss, BALB-c, CBA, C56Bl/6, DBA-2, and B6D2F1 strains) on days 0 and 2 of the experiment. On day 2, Freund's Complete Adjuvant was injected subcutaneously into the external ears of each mouse. On day 9, the thickness of the external ear was measured with a mobile-disk caliper immediately before and 24 h after a topical application of 100% PG to the treated external ears of each of the test mice. There was no significant increase in the external ear thickness, a measure of skin irritation and sensitization, in any of the test animals as compared with controls (Descotes, 1988). Clark et al. (1979) investigated the irritant effects of PG in rabbits. PG (con centration not stated) was applied to intact and abraded skin on the backs of six New Zealand white rabbits according to the methods of Draize et al. (1944). Irritation was scored at 24 and 72 h after treatment. The irritation score, obtained by a compilation of scores obtained at 24 and 72 h, was 0.1 (maximum possible = 8). A score of <2 was classified as a mild to no irritation reaction. PG was applied to the clipped skin (intact and scarified) of the backs of three groups of six New Zealand white rabbits according to three different protocols by Guillot et al. (1982b). In protocol! (the cosmetic protocol), rabbits were exposed to 0.5 ml of the test substance for 23 h under occlusive patches. Protocol 2 [the Association Francaise de Normalisation (AFNOR) protocol] was defined as the exposure of rabbits to 0.5 ml ofthe test substance for 4 h under occlusive patches. In protocol 3 [the Organisation for Economic Co-operation and Development (OECD) protocol], rabbits were exposed to 0.5 ml of the test substance for 4 h under semiocclusive patches. In each of the three protocols, both abraded and intact skin sites were tested. Reactions were scored at 1 and 48 h in the cosmetic protocol, 1, 24, and 48 h in the AFNOR protocol, and 30--60 min, 24, 48, and 72 h in the OECD protocol. The primary cutaneous irritation index obtained for the three protocols were 0.17, 0.14, and 0.02, respectively. In all three tests, PG was classified as a nonirritant. Open and closed patch tests were conducted by Motoyoshi et al. (1984) using guinea pigs and rabbit~. Patches containing 0.1 ml PG/patch (four per animal; two open, two closed) were applied for 48 h to the shaved skin of six Hartley guinea pigs. Patches containing 0.3 ml PG/patch (four per animal; two open, two closed) were applied to the shaved skin of six Angora rabbits for 48 h. Reactions were scored 30 min after removal of the patches. There were no reactions noted in either the open or the closed patch tests in either the rabbits or the guinea pigs. The authors noted that both of the species tested lack sweat glands, which might have accounted for the lack of a reaction. Forty-eight-hour and 21-day open and closed patch tests were conducted by these authors using miniature Gottingen swine. Four patches per animal (two open, two closed) contajning 0.1 ml PG/patch were applied to the shaved skin of two swine. Reactions were scored 30 min after the 48-h exposure. In the 21- day test, reactions were scored daily. Skin irritation reactions were not observed in any of the animals. The authors concluded that the previously mentioned lack

J A!11 Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 463 of sweat glands may have contributed to the lack of a reaction after exposure to PG. Pure PG was applied to the flanks of 27 guinea pigs and to the backs of 20 rabbits in a skin fold thickness test. Open patches containing the test material (amount unspecified) were applied daily to the test sites. Untreated sites served as controls. The skin fold thickness, a measurement of skin irritation, was measured daily with a Harpenden skin fold caliper. Sites were also examined for erythema, edema, and scaling. There were no significant increases in skin fold thickness in the rabbits. Most test sites in the rabbits had transient redness, which resolved spontaneously. There was a significant increase in skin fold thickness in the guinea pigs on day 7 (p < 0.01 and on days 8-10 (p < 0.001) (Wahlberg and Nilsson, 1984). Kero and Hannuksela (1980) tested the induction of contact hypersensitivity reactions by PG using three different tests. Using the guinea pig maximization test (GPMT), 20 animals were injected with 70% PG. In the open epicutaneous test (OET), 20 guinea pigs had open patches containing 70% PG applied to an 8-cm2 shaved area of skin during a 21-day period. In the chamber test (CT), eight Finn chambers containing 20 f.Ll 70% PG were applied to the shaved skin of each of 20 guinea pigs for 6 days. Chambers were changed every 48 h. Challenges were done on day 21 in the GPMT and the CT and on day 31 in the OET. At challenge, no positive reactions were noted in any of the animals in the three tests. PG did not induce contact hypersensitivity in these tests. Using the epicutaneous maximization test, Guillot and Gonnet (1985) investi gated the sensitization potential of PG in guinea pigs. Seven successive 48-h occlusive patches containing PG (amount not specified) were applied to the shaved skin of each of 20 Dunkin-Hartley guinea pigs. After a 1-week nontreat ment period, one 48-h occlusive challenge patch was applied to a previously untreated site on each guinea pig. Readings were taken at 1, 6, 24, and 48 h. Skin samples were taken for microscopic examination. No irritation or sensitization reactions were noted in any of the test animals. PG was neither an irritant nor a sensitizer in this test. Guillot et al. (1983) studied the sensitizing potential of several chemicals, in cluding PG, using seven different protocols. In the first protocol, the authors followed the methods of Magnusson and Kligman (1969) in a GPMT. On day 0 of the experiment, a 48-h occlusive patch containing 0.5 ml PG was applied to the dorsal surface of each of 20 Dunkin-Hartley guinea pigs. On day 7, the skin was painted with 0.5 ml 10% sodium lauryl sulfate. On day 8, a 48-h occlusive patch containing 0.5 ml PG was applied to the same area. After a 10-day nontreatment period, a 48-h occlusive challenge patch containing 0.5 ml PG was applied to the left flank of each animal. Reactions were scored 24 and 48 h after removal of the challenge patch. Skin samples were taken for microscopic examination. No re actions were noted in this test. PG was classified as a Grade I chemical in this test (Grade I = weak sensitizing potential). In the second protocol, the split adjuvant technique, the authors followed the methods of Maguire (1973). Three groups of guinea pigs, 10 per group (5 males and 5 females), had 0.1 ml PG applied under an occlusive patch after a 5-s treatment

JAm Col/ Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

464 COSMETIC INGREDIENT REVIEW

with dry ice on the dorsal surface (day 0 of the experiment). On day 2, a second 48-h occlusive patch was applied to the same area of the body. On day 4, animals received intradermal injections of 0.1 ml Freund's Complete Adjuvant, followed by a third 48-h occlusive patch containing 0.1 m1 PG. Mter a 12-day nontreatment period, all groups had 24-h-occlusive patches containing 0.1 m1 PG applied to the test areas on day 21 of the experiment. Group 1 animals were examined at 24, 48, and 72 h for reactions. Groups 2 and 3 received 24-h-occlusive patches containing the same amount of PG on day 35 of the experiment. Group 2 animals were examined at 24, 48, and 72 h for reactions. Group 3 animals received 24-h occlu sive patches containing 0.1 m1 PG on day 42 of the experiment and were examined 24, 48, and 72 h later for reactions. Results were scored on a five-point scale. PG was nonallergenic in this test. Guillot et al. (1983) followed the methods of Maurer et al. (1975) to test the sensitization potential of PG in a guinea pig optimization test. Two intradermal injections of 0.1 ml PG were administered to each of 20 guinea pigs (10 males and 10 females) on day 0 of the experiment. On days 2 and 4, single injections of the same concentration of PG were administered, followed by depilation of the test site 21 h later. Reactions at the test sites were scored after depilation. On days 7, 9, 11 14, 16, and 18, animals received single intradermal injections of 50% PG/50% Freund's Complete Adjuvant. Following a 2-week nontreatment period, animals were depilated and received one intradermal injection of 0.1 ml PG on day 35. Test site reactions were scored after depilation. On day 49, 24-h occlusive patches containing 0.1 ml PG were applied to the test sites. On day 51, following the removal of the patches and depilation of the test site, irritation/sensitization re actions were evaluated. No reactions were noted in any of the test animals. PG was nonallergenic in this test. These authors also investigated the irritation/sensitization potential of PG using a method outlined by Guillot and Brulos (unpublished protocol). In the induction phase of the experiment, 20 guinea pigs (10 male and 10 female) were each given single intradermal injections of 0.1 ml Freund's Complete Adjuvant, followed by 48-h occlusive patches containing 0.5 m1 PG. On days 2, 4, 7, 9, 11, and 14, 49-h occlusive patches were reapplied. Following a 22-day nontreatment period, 48-h occlusive patches containing 0.5 ml PG were applied in the challenge phase of the experiment. At 1, 6, 24, and 48 h after removal of the challenge patches, macro scopic and histological examinations of the test sites were conducted. There were no reactions in any of the test animals, and the test score was 0 (0 = sensitizing potential weak or nonsignificant). In the fifth protocol (Freund's Complete Adjuvant test), Guillot et al. (1983) followed the methods of Klecak et al. (1977). Two groups of 10 Dunkin-Hartley guinea pigs (5 male/5 female per group) received intradermal injections of 0.1 m1 PG on days 0, 2, 4, 7, and 9 of the induction phase of the experiment. Following a 10-day nontreatment period, both groups received topical applications ofPG (25 JLl), and reactions were scored at 24, 48, and 72 h. Group 2 animals received an additional challenge of 25 JLl PG on day 35, and reactions were scored at 24, 48, and 72 h. No reactions were noted. PG was nonallergenic in this test. In the sixth test protocol, they followed the methods of Dossou and Sicard

JAm CoU Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 465

(1984). Two groups of guinea pigs, 12 per group, were used to test the irritation/ sensitization potential ofPG. Group 1 animals received 0.5-ml dermal applications of PG in an open test on days 0, 2, and 4 of the experiment. Following a 10-day nontreatment period, challenge applications of 10 JJ..l PG were applied to the skin. Reactions were scored 24 and 48 h after the challenge applications. Group 2 animals received injections of 50% PG/50% Freund's Complete Adjuvant on day 4 of the experiment. Following a 5-day nontreatment period, challenge applica tions of 10 JJ..l PG were made to the skin. Reactions were scored at 24 and 48 h. No reactions were noted in either of the test groups. PG was nonallergenic in this test. In the last protocol tested by Guillot et al. (1983), an OET was performed according to the methods of Klecak et al. (1977). Two groups of guinea pigs, eight per group, received single topical applications of 0.1 ml PG on the right posterior flank on days 0-20 of the experiment. On day 21, challenge applications of 25 JJ..l PG were made to the test sites. Group 1 animals were examined at 24, 48, and 72 h for reactions. Following a 14-day nontreatment period, group 2 animals received additional challenge exposures of 25 JJ..l PG per animal. Macroscopic and histo logical examinations were performed 24, 48, and 72 h later. No reactions were noted in any of the animals in either test group. PG was nonallergenic in this test. Jacaruso et al. (1985) investigated the irritation potential of PG in an in vitro histamine release study. Cells were harvested from male Wistar rats, and a pop ulation containing mast cells, macrophages, leukocytes, and erythrocytes was isolated and suspended in HBSS and bovine serum albumin (BSA). PG, at final concentrations of0.01, 0.10, 1.0, or 10%, was added to the cells for 10 min (37°C). The reaction mixture was spun at 400 g for 10 min, and the supernatant was used for analysis of histamine release, an index of irritation potential. Control values for histamine release were obtained by exposing a cell population to HBSS/BSA alone. Only those cells exposed to a final concentration of 1.0% PG released significantly more histamine than control cells (p < 0.01).

Polypropylene Glycols

Skin irritation was not noted after PPG 425, PPG 1025, or PPG 2025 when applied once to clipped skin (abdomen) of the rabbit or when applied a total of eight times to the same area within 4 h. Details concerning the experimental procedure were not stated (Shaffer et al., 1951). The skin irritation potential of undiluted PPG 1200 was evaluated using an albino rabbit (weight 2.3 kg). The test substance (1-2 ml) was applied to intact and abraded abdominal skin by means of an absorbent cotton pad held in place by a cloth bandage that was taped to the trunk. Prolonged applications were made to intact skin (daily for 3 days) and abraded skin (5 days/week for 2 weeks). Repeated applications (5 days/week for 2 weeks) were also made to the inner surface of the external ear. Reactions were not observed at the intact skin site. Barely percep tible erythema and slight exfoliation were observed at the abraded site. There was no perceivable skin damage after multiple applications to the inner surface of the external ear (FDA, 1992).

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

466 COSMETIC INGREDIENT REVIEW

The results of skin irritation and sensitization studies on PG are summarized in Table 9. Studies on the skin irritation potential of PPG are summarized in Ta ble 10.

Hematotoxicity The hemolytic effects of PG were evaluated in a 117-day study using 21 adult, specific-pathogen-free cats (7 males, 14 females; weights between 2.3 and 5.1 kg). The animals were randomly divided into three groups and fed 6% PG, 12% PG (in

TABLE 9. Irritation/sensitization potential of Propylene Glycol

Dose/concentration; Animal No. exposure method Results Ref. Nude mice 3 10, 25, 50%; 24-h exposure in No irritation Lashmar et al. (1989) PVC cup Mice 19 I 00%; mouse ear sensitization No irritation/ Descotes (1988) assay sensitization Rabbits 6 Unknown concentration; 24-h Mild to no irritation Clark et al. (1979) patch test Rabbits 6 Cosmetic protocol (see text) N onirritant Guillot eta!. (1982b) Rabbits 6 AFNOR protocol (see text) N onirritant Guillot et al. (1982b) Rabbits 6 OECD protocol (see text) Nonirritant Guillot et al. (1982b) Guinea pigs 6 0.1 ml; 48-h open patch test No observed reaction Motoyoshi et al. (1984) Guinea pigs 6 0.1 ml; 48-h closed patch test No observed reaction Motoyoshi et al. (1984) Rabbits 6 0.1 ml; 48-h open patch test No observed reaction Motoyoshi et al. (1984) Rabbits 6 0.1 ml; 48-h closed patch test No observed reaction Motoyoshi eta!. (1984) Swine 2 0.1 ml; 48-h open patch test No skin irritation Motoyoshi et al. (1984) Swine 2 0.1 ml; 48-h closed patch test No skin irritation Motoyoshi et al. (1984) Swine 2 0.1 ml; 21-day open patch test No skin irritation Motoyoshi et a!. (1984) Swine 2 0.1 ml; 21-day closed patch No skin irritation Motoyoshi et a!. (1984) test Guinea pigs 27 Daily open patches, skin fold Days 7 and 8-10, significant Wahlberg and Nilsson thickness assay increase in skin fold (1984) thickness Rabbits 20 Daily open patches, skin fold No increase in skinfold Wahlberg and Nilsson, thickness assay thickness; transient redness 1984 Guinea pigs 20 70%; maximization test No observed reaction Kero and Hannuksela (1980) Guinea pigs 20 70%; open epicutaneous test No observed reaction Kero and Hannuksela (1980) Guinea pigs 20 20 f.'l 70%; three 48-h No observed reaction Kero and Hannuksela chamber exposures (1980) Guinea pigs 20 Epicutaneous maximization No irritation/ Guillot and Gonnet (1985) test sensitization Guinea pigs 20 0.1 ml; protocol I, Weak irritation/ Guillot et al. (1983) maximization test sensitization potential Guinea pigs 30 0.1 ml; protocol2, split Nonallergic Guillot et al. (1983) adjuvant technique Guinea pigs 20 0.1 ml; protocol 3, Nonallergic Guillot et a!. (1983) optimization test Guinea pigs 20 0.5 ml; protocol 4, Nonsignificant or weak Guillot et al. (1983) Guillot/Brulos sensitizer Guinea pigs 20 0.1 ml; protocol 5, Freund's Nonallergenic Guillot et a!. (1983) Adjuvant test Guinea pigs 24 Protocol 6 (see text) Nonallergenic Guillot et al. (1983) Guinea pigs 16 0.1 ml; protocol 7, open Nonallergenic Guillot et a!. (1983) epicutaneous test Rats O.oJ, 0.10, 1.0, and 10%; in Low skin irritation potential Jacaruso et al. (1985) vitro histamine release assay

For abbreviations and details, see the text.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 467

TABLE 10. Skin irritation potential of Polypropylene Glycol Dose/concentration; Animals No. exposure method Results Ref. Rabbits Single application and 8 No skin irritation Shaffer et al. (1951) applications within 4 h Rabbits Single application and 8 No skin irritation Shaffer et a!. (1951) applications within 4 h Rabbits Single application and 8 No skin irritation Shaffer et a!. (1951) applications within 4 h Rabbit Applications to intact skin No reactions at intact site; FDA (1992) daily for 3 days; barely perceptible applications to abraded erythema and slight skin 5 days/wk for 2 exfoliation at abraded wks site Rabbit Applications to different No perceivable skin FDA (1992) intact skin site 5 damage days/wk for 2 wks

the diet), and a control diet, respectively. Blood was drawn at the beginning of the study and every 2 weeks thereafter. Reticulocyte and Heinz body counts were expressed as the percentage of 1,000 cells counted. Differences between treat ment groups and changes over time within groups were evaluated by an analysis of variance. In experimental and control groups, the packed cell volume de creased over time and was significantly lower in the 12% group only during the sixth week of the study. Significant changes in hemoglobin concentration were not observed between groups; however, a statistically significant (p < 0.001) decrease over time was observed in the 6% PG group. Throughout the study, a significant decrease (p < 0.001) in numbers of erythrocytes was noted in 6 and 12% PG groups, but not in the control group. A significant (p < 0.0001) increase in punc tate reticulocytes over time was observed in 6 and 12% PG groups, but not in the control group. Additionally, a significant increase in the number of aggregate reticulocytes was noted in the 12% PG group 2 weeks after initiation of PG containing diets. Heinz body counts were significantly higher, in a dose-related manner, in 6 and 12% PG groups; the increase persisted throughout the study. The results of this study indicated that PG-containing diets caused a dose-dependent erythrocyte destruction, even when fed at concentrations as low as 6% (Bauer et al., 1992). Decreased erythrocyte survival and Heinz body formation were noted · in an earlier study in which three cats (weights 3-4 kg) were fed a dosage of 8 g PG/kg body wt daily for 22 days (Christopher et al., 1990b). Heinz body formation was also noted in groups of six kittens (weight 2,160 ± 47 g) fed diets containing 5 and 10% PG for 12 weeks. None of the animals had anemia or methemoglobi nemia (Hickman et al., 1990).

Reproductive Effects A continuous breeding reproduction study was conducted using COBS Crl:CD-1 (ICR)BR outbred Swiss albino mice (6 weeks old). The continuous breeding phase of the study (task II) was begun after the dose-setting study (task

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

468 COSMETIC INGREDIENT REVIEW

I) and involved three experimental groups (40 mice per group) and a control group of 80 mice. Experimental and control groups contained an equal number of male and female mice. The three experimental groups were given the following doses (in feed or water), respectively, during a 7-day premating period: 1.0% PG (daily dose 1.82 g/kg), 2.5% PG (daily dose 4.80 g/kg), and 5.0% PG (daily dose 10.10 g/kg). The animals were then randomly grouped as mating pairs, cohabitated, and treated continuously for 98 days; data were collected on all newborns. If signifi cant adverse effects on fertility were observed, a crossover mating trial (task III) was usually performed to determine whether F0 males or females were more sensitive to the effects. Task III was not conducted and would have consisted of mating high-dose mice of each sex to control mice of the opposite sex and then analyzing the offspring. To perform an offspring assessment of reproductive func tion (task IV) following exposure to PG, the dam (from phase II) was dosed through weaning and F 1 mice were dosed until mating occurred at 74 ± 10 days of age. Mating pairs consisted of male and female offspring from the same treatment group (20/group/sex); F 2 litters were examined. In the continuous breeding phase (task II), there were no significant changes (p < 0.05) in mean live pup weight per litter between the control group and any of the treatment groups. In task IV, which was an offspring assessment of reproductive function, only the high-dose group (5% PG) was involved. There were no significant differences (p < 0.05) between control and experimental groups with respect to the following observa tions in task IV: mating index, fertility index, mean number oflive pups per litter, proportion of pups born alive, and sex of pups born alive (Morrissey et al., 1989).

Embryotoxicity

The effect of PG on the development of B6D2F 1 mouse zygotes in the pronu clear stage was evaluated; oocytes were fertilized in vitro. Samples of zygotes were incubated for 20 min (at 22°C) with 1.5, 3.0, and 6.0 M PG in phosphate buffered saline. There were three zygote cultures per test concentration of PG, and each group of three were incubated with 0, 0.1, and 0.25 M sucrose, respec tively. The three control cultures without PG were also incubated with 0, 0.1, and 0.25 M sucrose, respectively. Subsequently, the zygotes were incubated for 24 h (at 37°C) under 5% C02, and the percentage of zygotes that cleaved to form two-cell embryos was determined. The percentage of zygotes that developed to two-cell embryos was not altered in control cultures or cultures exposed to 1.5 M PG (78% in both cultures), but was reduced in cultures exposed to 3.0 M PG (7%; p < 0.05). Embryonic development was inhibited completely in zygotes exposed to 6.0 M PG. The presence of sucrose in the incubation medium did not influence embryonic development (Damien et al., 1989). In a later publication by the same authors, the data indicated that the exposure of B6D2F 1 mouse zygotes to ~2.5 M PG for 2-7 min altered both intracellular pH and developmental potential. In that these effects were independent of volume changes noted in zygotes and therefore intracellular PG concentrations, the authors postulated that the toxicity of PG is mediated by direct alteration of the cell membrane (Damien et al., 1990).

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 469

Teratogenicity Kavlock et al. (1987) investigated the teratogenic potential of several sub stances, including PG. A group of 30 pregnant female CD-1 mice was given single oral doses of 10,000 ppm PG on days 8-12 of gestation. Fertility rates, numbers of maternal deaths, numbers of resorptions, average litter sizes, birth weights, and pup postnatal weight gain were monitored in an assessment of the maternal and perinatal effects of PG. The fertility rates of mice given PG were not significantly different from those of control mice. There were no maternal deaths or resorp tions observed in any of the animals dosed with PG. All other parameters mea sured were not significantly different from control values. PG was not a teratogen in this test. PG (dose 0.01 mUg body wt) also was injected subcutaneously into each of 21 pregnant ICR/Jcl female mice (9-12 weeks old) on day 9, 10, or 11 of gestation. Days 9-11 correspond to the sensitive stage for the induction of fetal deaths and malformations in this strain of mice. The following malformations were noted in 5 of the 226living fetuses: open eyelid (3 fetuses), polydactyly (1 fetus), and cleft palate (1 fetus). In a control group of 28 mice (same strain and weight range) injected subcutaneously with water (0.01 rnVg body wt) during pregnancy, the only malformation noted was exencephalus in 1 of 320 living fetuses. The inci dence of malformations in 1,026 living fetuses from an untreated control group of 90 pregnant mice (same strain and weight range) was 3 fetuses with polydactyly, exencephalus, and open eyelid, respectively (Nomura, 1977). Ascitic mouse ovarian tumor cells were used in an in vitro teratogenicity assay by Braun et al. (1982). Tumor cells were labeled with eHJthymidine in situ via an intraperitoneal injection of 0.2 mCi of radioactivity. Cells were harvested and suspended in phosphate-buffered saline (107 cells/ml). Various concentrations of PG were added to aliquots of the cells (all concentrations not stated) and incu bated at 37°C for 3o min. The teratogenicity of the substances tested was assayed by the ability of the test substances to inhibit attachment of the tumor cells to Concanavalin A-coated plastic. The extent of attachment of the cells was mea sured by counting the radioactivity on the Concanavalin A-coated plastic. PG did not inhibit attachment of the tumor cells; the largest dose tested was 27,000 mg/L PG. PG was a nonteratogen in this test.

Mutagenicity In Vitro Tests Clark et al. (1979) assayed the mutagenicity of heat transfer fluids, including PG, according to the methods of Ames et al. (1975). PG was diluted with dimeth ylsulfoxide (DMSO) to final concentrations of 1-10,000 !J.g/plate (actual concen trations tested not stated). PG was added to tester strains TA 1535, TA 1537, TA 1538, TA 98, and TA 100 both with and without metabolic activation. PG was not mutagenic in any of the strains at any concentration tested. PG in DMSO was tested in a reverse mutation assay (Ames et al., 1975) in volving Salmonella typhimurium strains TA 92, TA 1535, TA 100, TA 1537, TA 94, and TA 98. The dose tested was 10 mg/plate, largest noncytotoxic dose, and cul-

JAm Colt Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

470 COSMETIC INGREDIENT REVIEW

tures were incubated in the presence of metabolic activation. Results were con sidered positive if the number of revertant colonies in experimental cultures was twice that noted in the solvent control culture. PG was not mutagenic in this assay (lshidate et al., 1984). Sasaki et al. (1980) investigated the mutagenic effects of 60 chemicals, including PG, using an HE 2144 human fibroblast cell line and a Don-6 Chinese hamster cell line. Concentrations of 3.805, 7.61, or 22.83 mg/ml PG in HBSS, added to 1.0 J.Lg 5-bromodeoxyuridine (BrdU), were incubated for two cell cycles with either the human or the Chinese hamster cell line. Chromosome slides were stained with acridine orange fluorescent stain, and the frequencies of sister chromatid ex changes (SCEs) were scored. The authors reported a dose-dependent increase in the frequencies of SCEs in the Chinese hamster cell line. PG was a weak, but potential, SCE inducer in this investigation. The mutagenicity of PG was also evaluated in a chromosomal aberrations test using a Chinese hamster fibroblast cell line. The cells were exposed to 32.0 mg/ml PG in physiological saline (maximum dose) for 48 h without metabolic activation. The maximum dose was defined as the dose at which the maximum effect was obtained. Untreated cells and solvent-treated cells served as controls; the inci dence of aberrations was usually <3.0% in these cells. The incidence of polyploid cells and cells with structural chromosomal aberrations, such as chromatid or chromosome gaps, breaks, exchanges, ring formations, fragmentations, and oth ers, were recorded for each culture. The results were considered positive if the total incidence of cells with aberrations (including gaps) was ~10%, equivocal if the incidence was between 5 and 9.9% and negative if the incidence was ~4.9%. Positive results were reported for PG in this test (Ishidate et al., 1984). Abe and Sasaki (1982) investigated the mutagenicity of PG in vitro in an SCE assay. PG was tested in both human cultured fibroblasts and a cultured Chinese hamster cell line, with and without the addition of liver S9 mix as a metabolic activator. SCE was measured after the addition of BrdU. PG was not mutagenic in this assay system. PG was tested in an in vitro DNA damage alkaline elution assay according to a modification of the procedure by Kohn et al. (1974). The procedure involved incubation of Chinese hamster lung fibroblast V79 cells, in the presence and absence of metabolic activation, with 10 mM PG for 1, 2, and 4 h. Prior to PG exposure, DNA in the cells was labeled with e4C]thymidine. Single-stranded DNA was eluted using a single polyvinyl filter, and the recovery of radioactive DNA applied to the filter was >95%. DNA damage was not observed in PG exposed cultures (Swenberg et al., 1976). PG was not mutagenic in other short-term in vitro tests: chromosomal aberra tions (in anaphase) in human embryonic lung cells (WI-38), mitotic recombination in Saccharomyces cerevisiae strain D4, and basepair substitution in Salmonella typhimurium strains G-46 and TA-1530 (Green, 1977).

In Vivo Test The mutagenicity of PG was evaluated in the micronucleus test using 25 (five groups of 5) 8-week-old ddY mice. Single intraperitoneal injections of PG were

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 471 administered to four groups in doses of 2,500, 5,000, 10,000, and 15,000 mg/kg in saline, respectively. The control group was dosed with saline. Femoral marrow cells were smeared on glass slides, 1,000 polychromatic erythrocytes per mouse were scored, and the number of micronucleated polychromatic erythrocytes was recorded. Results were classified as positive only when a statistically significant difference was noted between one or more treatment group(s) and the spontane ous incidence of micronucleated polychromatic erythrocytes, and the Cochran Armitage trend test (Armitage, 1955; Cochran, 1954) indicated a positive dose response. PG was not mutagenic at any of the doses tested (Hayashi et al., 1988).

Carcinogenicity In Vitro Test Pienta (1980) investigated the carcinogenic potential of PG in a hamster embryo cell transformation bioassay. Embryo cells were isolated from Syrian hamsters on day 13 of gestation. Isolated cells were separated into tubes containing 2.5 x 106 cells and frozen at - 195°C prior to the experiment. Samples of the frozen cell aliquots were reconstituted on day 0 of the experiment and cultured for 6 days. On day 6 of the experiment 4 ml of PG at concentrations ranging from 0.125 to 8.0% (actual concentrations not stated) was added to the cells. Positive and negative controls were tested in each experiment. After 8 days, the cultures were investi gated for the presence of any transformed colonies. No transformed colonies were reported in any of the PG-dosed cells.

In Vivo Tests Gaunt et al. (1972) investigated the carcinogenicity of PG in a 2-year feeding study using CD strain rats (methods of study previously described in Chronic Oral Toxicity). The types and numbers of neoplasms found were documented during the study. In female rats, the largest numbers of neoplasms found were mammary gland fibroadenomas. The authors indicated that these neoplasms occurred fre quently in rats of the CD strain. There were no significant differences in the incidence of neoplasms found in the PG-treated animals as compared with un treated control animals. There was no observed dose-response relationship be tween animals fed low and high concentrations of PG. The authors reported that no carcinogenic potential was found when concentrations up to 50,000 ppm PG were administered in the diet. Stenback and Shubik (1974) investigated the carcinogenicity of PG using Swiss mice. Three groups of animals, 50 per group, had concentrations of 10, 50, or 100% PG (diluted in acetone where necessary) applied to the shaved dorsal skin between the flanks twice a week. Applications continued for the lifetime of the animals. All lesions and neoplasms that developed were recorded weekly during the experiment; necropsy was performed on all animals. In the group that re ceived 10% PG, 26 of 50 animals (52%) developed neoplasms; however, skin tumors were not observed. In the 50 and 100% groups, the numbers of tumor bearing animals were 26 of 50 (52%) and 20 of 50 (40%), respectively. The distri-

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

472 COSMETIC INGREDIENT REVIEW bution of the more prevalent tumors in the three treatment groups was as follows: 10% PG (15lymphomas, 7lung adenomas, 1liver hemangioma, and 1 thymoma), 50% PG (13 lymphomas, 13 lung adenomas, 2 liver hemangiomas, and 2 thymo mas), and 100% PG (9lymphomas, 7 lung adenomas, 4 liver hemangiomas, and 1 thymoma)~ In vehicle (50 mice, 100% acetone) and positive control (0.5% 7,12- Dimethylbenzanthracene, 50 mice) groups, the percentages of tumor-bearing an imals were 44 and 78%, respectively. The percentage of tumor-bearing animals in the untreated control group (150 mice) was 42%. The authors reported that there was no significant carcinogenic potential when PG concentrations up to 100% were applied to the skin of Swiss mice. PG was not found to be carcinogenic in other carcinogenicity studies, whether the method of administration was oral, cutaneous surface application, or subcu taneous injection (Fujino et al., 1965; Dewhurst et al., 1972; Wallenius and Lekholm, 1973a,b; du Vivier and Stoughton, 1975; Farsund, 1978).

CLINICAL ASSESSMENT OF SAFETY Short-Term Dermal Toxicity

An aqueous cream containing 60% PG was applied daily to seven inpatients with psoriasis (one patient was diabetic) for 5 days; doses ranged from 1.5 to 6.1 g/kg/24 h (Commens, 1990). It has been calculated that a 70-kg adult would re quire, in normal practice, ~44 g of cream for total body coverage twice per day, resulting in an approximate dosage of0.35 g/kg. Total-body occlusion therapy was abandoned because three of the patients were eliminated from the study due to skin irritation. Blood was collected at the beginning of application, on day 3, and at the end of the application period for measurement of the following: serum lactate, urea, electrolytes, glucose, and osmolality levels. Two of the remaining four patients had mild renal impairment, and most patients experienced desqua mation either during or after the study. No significant changes in serum osmolality and lactate were noted in any of the patients. Hyperosmolality has been induced by PG in a number of clinical settings (Glasgow et al. 1983), particularly in inten sive care unit patients during the administration of nitroglycerin solutions that contain PG (Bossaert and Demey, 1987; Demey and Bossaert, 1987). Addition ally, hyperosmolality has resulted from the percutaneous absorption of PG in bum patients (Fligner and Jack, 1985; Kulick et al., 1980; Bekeris et al., 1979).

Skin Irritation and Sensitization

In a review article by Catanzaro and Smith (1991), it was noted that a commonly encountered problem in patch testing, particularly with PG, is reliably differenti ating an irritant reaction from an allergic reaction. This is true especially if the reaction is relatively weak. Accordingly, several authors have been unable to conclude whether patch test reactions observed in subjects represented irritation or true allergic sensitization (Blondeel et al., 1978; Romaguera et al., 1981; Han nuksela and Salo, 1986; Kinnunen and Kannuksela, 1989). These studies and the

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 473 results of other human skin irritation and sensitization tests are included in this section. Predictive Tests Propylene Glycol. Wahlberg and Nilsson (1984) studied the irritant effects ofPG in humans (number of test subjects not indicated) using laser Doppler flowmetry. PG was applied using three different methods: single open exposures to 1.0 m1 PG on the ventral aspect of the thigh with and without occlusive patches for 5-15 min and repeated open exposure to 1.0 ml PG on the skin for 12 days. Blood flow at the test site was measured and used as an indication of irritation. PG, when administered under occlusive patches, caused weak erythema at the test site; the maximum reaction was measured 26 h after exposure. Single open applications and repeated open exposure to PG did not cause any irritation reactions in this test. These authors also used a skin fold thickness test to further study the irritant effects of PG in humans (number of test subjects not mentioned). Concentrations of 100% PG were applied to the volar forearm of the test subjects for 36 consec utive days. Skin fold measurements were made daily using a Harpenden skin fold caliper. Unexposed skin sites were used as controls. Under the conditions of this test, PG did not cause any increase in skin fold thickness in any of the test subjects. Willis et al. (1989) studied the irritant effects of PG in 48-h patch tests. Ten adult males received patches, containing 100% PG, in Finn chambers applied to the volar aspect of the forearm for 48 h. The test sites were examined for any reaction 1 h after removal of the patches. Reactions were characterized as either negative, mild, moderate, or severe. Skin biopsies were also performed, and samples ob tained were examined microscopically. Reactions were not observed when the application sites were examined macroscopically. Upon microscopic examina tion, half of the samples contained occasional areas of slight spongiosis, accom panied by a minimal infiltrate of mononuclear cells into the epidermis. No signif icant irritant effects were reported in this test. Motoyoshi et al. (1984) conducted both 48-h J?atch tests and 21-day repeated insult patch tests to assess the irritation potential of PG in humans. In the 48-h patch tests, patches containing 0.05 mllOO% PG were applied to the backs of 50 adult men. After patches were removed, the sites were scored for irritation and skin samples were taken for microscopic examination. Exposure to PG produced erythema and edema in a dose-dependent manner in the test subjects (number of reactions not specified). In 21-day patch tests by these authors, 24 adult men received patches contain ing 0.05 ml of 1, 3, 10, or 30% PG. Patches were applied daily to the upper backs of each of the test subjects. Sites were scored for irritation and skin samples were taken for microscopic examination. Concentrations of 10 and 30% PG produced irritation reactions in an unspecified number of subjects. The authors concluded that PG caused primary irritation reactions in this test. Ten healthy nonatopic male volunteers were patch tested with 100% PG. The test substance was applied for 48 h to the volar aspect of the forearm using 8-mm

JAm Coil Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

474 COSMETIC INGREDIENT REVIEW

Finn chambers, and reactions were scored 1 h after patch removal according to the following scale: 0 (negative) to 3 (erythema, edema, and vesiculation). Half of the subjects were patch tested with distilled water, negative control, under oc clusive patches. Punch biopsies were obtained from each patch test site, and skin samples were processed for immunocytochemistry and electron microscopy. Concerning the immunocytochemical technique, the antibody OKT6, which spe cifically recognizes the CD1a antigen expressed on the surface of the Langerhans cells and indeterminate dendritic cells in the skin, was used. The Student paired t test was used to compare the experimental CD1 + cell density and CD1 + cell dendrite length with that of the control. The majority of the reactions observed were mild to moderate, and the number and dendrite length of CD1 + cells within the epidermis at PG-treated sites were not significantly different from the obser vations at negative control sites. The Spearman rank coefficient of correlation indicated a nonlinear correlation between CD1 + cell density and clinical grading (r = 0.1157). However, there appeared to have been a relationship between these parameters. Mild reactions generally caused a slight decrease in density, and severe reactions caused an increase in CD1 +cell number. As for dendrite length, the greater the intensity of the reaction, the shorter the dendrite (Willis et al., 1990). Willis et al. (1988) performed an experiment to determine optimum concentra tions of several compounds, including PG, which would induce irritation in 75% of a test population. Two concentrations of PG, 50 and 100%, were tested using 16 and 35 healthy volunteers, respectively. Each subject received an application of PG, under occlusive patches for 48 h, on the volar area of the forearm via a Finn chamber. At the end of the exposure period, the patches were removed and the test sites were scored 1 h later. The grading system used was - (no reaction) to 4 + (intense erythema with bullous formation). In the subjects tested with 50% PG, there were no positive reactions. Among the subjects tested with 100% PG, 14 (40%) positive reactions were reported (11 ± reactions, 3 1 + reactions). PG had marginal irritant properties in this test. The contact sensitization potential of 12% PG in a cream vehicle was tested using a panel of 204 test subjects. In the induction phase of the experiment, 10 successive 48- or 72-h patches containing 0.5 g PG were applied under occlusive patches to the upper arms of each of the test subjects. Following a 2-week non treatment period, a 72-h occlusive challenge patch containing 0.5 g PG was ap plied to the test site on each of the panelists. Reactions were scored after removal of the challenge patch. There were no reactions noted in any of the 204 panelists in this experiment (Marzulli and Maibach, 1973). A 21-day cumulative irritancy assay was conducted by Patel et al. (1985) using PG on healthy volunteers. Webril patches containing PG (concentration not stated) were applied under occlusive patches to the back of each of 25 subjects. Patches were removed and changed daily for a total of 21 days. Test sites were examined for irritation daily prior to application of the next patch. The scale used for scoring ranged from 0 (no dermatitis) to 4 (erythema, induration, vesicles, and/or bullae covering the test site). The total cumulative irritation score was 72.0 in this test.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 475

Trancik and Maibach (1982) investigated the irritation/sensitization potential of PG in three tests. In a cumulative irritation procedure, 24-h occlusive patches containing 0.2 ml PG were applied to the backs of 10 subjects. A total of 21 consecutive patches were applied. Test sites were observed for reactions daily. In this 21-day test, one subject had an equivocal reaction on day 20. No other reactions were noted. These authors continued their investigation of the irritation/sensitization poten tial of PG in a modified Draize sensitization test. In the induction phase, a panel of 203 healthy volunteers had 48-h occlusive patches containing 0.2 ml PG applied to their backs. A total of 10 applications were made over a 21-day period. After a 2-week nontreatment period, challenge patches containing 0.2 ml PG were applied to fresh test sites and removed 48-72 h later. Reactions were scored on a scale of 0.0-4.0. During the induction phase, equivocal reactions (score = 0.5) were noted in eight of the panelists. Upon challenge, 19 cutaneous reactions were noted (6 had 0.5 reactions, 6 had 1.0 reactions, and 7 had 2.0 reactions). The same authors also conducted a provocative use test with PG using the 19 panelists with cutaneous reactions in the sensitization test described. Applications of 0.1 ml PG were made twice daily to the cubital fossa of each subject for a total of 7 days. Sites were scored for reactions on day 8. There were no reactions noted in any of the panelists. It was concluded that the cutaneous reactions observed in 19 subjects were irritant reactions. The results of predictive clinical tests on PG are summarized in Table 11. Polypropylene Glycol. A total of 300 human subjects received continuous and repeated dermal applications of undiluted PPG 2000. Details concerning the ex perimental procedure were not included. The test substance caused neither skin irritation nor sensitization (American Industrial Hygiene Association, 1980).

Provocative Tests

A total of 866 patients with various dermatological conditions were patch tested (closed or covered patches) with 100% PG from April 1951 to April 1952. The patches were applied to clinically normal skin, and test sites were examined 48 h after patch application. Positive reactions were observed in 138 (15.7%) patients. Reactions ranged from simple erythema ( +) to erythema with induration and vesiculation ( + + + + ). Eighty-nine of the 138 patients with positive reactions suffered from dermatitis venenata. A seasonal fluctuation in the incidence of positive reactions was also noted. The incidence was at a minimum when the climate was hot and humid (July, August, and September 1951 in New York City) and significantly greater during the cooler and less humid seasons. Of the 84 patients involved with simultaneous testing with several samples of PG from different sources, positive reactions were observed in 15. There were no differ ences in patient responses to different brands of PG (Warshaw and Herrmann, 1952). Twenty-three of the 138 patients with positive reactions to 100% PG in the preceding study were patch tested with aqueous PG. There were only five positive

JAm Col/ Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

476 COSMETIC INGREDIENT REVIEW

TABLE 11. Clinical predictive patch testing to evaluate skin irritation and sensitization potential of Propylene Glycol (PG)

Test No. of substance subjects Procedure Results Ref. 100% PG Not stated Single 5- to 15-min Weak erythema at test Wahlberg and Nilsson exposures to 1 ml of site (under (1984) PG (open and under occlusion); no occlusion) reactions (open) 100%PG 10 48-h patch test using No significant skin Willis et al. (1989) Finn chamber irritation 100%PG 50 48-h patch test; Erythema and edema Motoyoshi et al. 0.05-ml application noted (1984) 100%PG 10 48-h patch test using Mild to moderate Willis et al. (1990) Finn chamber erythema predominated 100% PG Not stated Repeated open No irritation Wahlberg and Nilsson exposures to 1 m1 (1984) PG for 12 days 100% PG Not stated Repeated open No increase in Wahlberg and Nilsson exposures for 36 skinfold thickness (1984) consecutive days PG(50 16 (50% 48-h exposure, under No reactions to 50% Willis et al. (1988) and 100%) PG group); occlusive patches PG; 11 ± and three 36 (100% 1 + reactions to group) 100%PG PG (1, 3, 10 24 Repeated closed patch Primary irritation with Motoyoshi et al. and 30%) applications for 21 10 and 30% PG (1984) days 12% PG in 204 10 successive 48- or No reactions Marzulli and Maibach cream vehicle 72-h occlusive (1973) patches (induction); 72-h occlusive challenge patch PG 25 Repeated exposures, Total cumulative Patel et al. (1985) under occlusive irritation score of 72 patches, for 21 days PG 10 Repeated exposures, 1 subject with Trancik and Maibach 0.2 ml under equivocal reaction (1982) occlusive patches, for 21 days PG 203 Modified Draize 13 positive reactions; Trancik and Maibach sensitization test; 10 6 equivocal (1982) 0.2-ml applications reactions under occlusive patches (induction); 48- to 72-h challenge PG 19 (subjects Provocative use test; Cutaneous reactions Trancik and Maibach with positive 0.1-ml applications noted in preceding (1982) reactions in twice daily for 7 study confirmed as preceding study) days irritant reactions

responses to 10% PG, and the application of 2.5% PG in water to 3 of the 23 patients resulted in one positive reaction. Additionally, 16 of the patients with positive reactions to 100% PG were also tested by simple inunction of the test substance. There was no evidence of an inflammatory response to the inunction either shortly after application or 48 h later (Warshaw and Herrmann, 1952). When 1,556 patients were patch tested with 100% PG, positive reactions were observed in 194 subjects. Four patients had "true allergy" and the remainder had irritant reactions. Three groups of 42 patients with positive reactions to 100% PG

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 477 were later tested with 3.2, 10, and 32% PG, respectively, and the results were as follows: 3.2% PG (9 positive reactions), 10% PG (12 positive reactions), and 32% PG (20 positive reactions) (Hannuksela et al., 1975; Catanzaro and Smith, 1991). Details concerning the experimental procedure were not included in the review article by Catanzaro and Smith (1991). The irritation/sensitization effects of PG were investigated using a panel of 38 patients with histories of PG reactions. Finn chambers containing 2, 10, 32, or 100% PG were applied to the back of each of the test subjects. Exposure lasted from 20 to 24 h, and reactions were scored, 1, 2, 4, and 5 days postexposure. Peroral challenge doses of 2% PG, either 2 or 5 ml, were administered to each of the test subjects. Reactions were recorded for 24 h or until compound-related signs had resolved. In the initiation phase of the experiment, allergic eczema was reported in 11 of 38 patients. Other types of eczema were noted in 16 of 38 patients. In 15 of the subjects who reacted to PG in the initiation phase, peroral challenge resulted in extensive exanthema, which resolved spontaneously within 48 h after challenge. The authors calculated that, based on the results of this experiment, 1% of all patients with eczema "may suffer harmful reactions either from the internal or external use of PG" (Hannuksela and Forstrom, 1978). In another study, positive reactions were observed in 12 of 84 patients patch tested with 100% PG. Five subjects had allergic reactions, and seven had skin irritation reactions (Andersen and Storrs, 1982). Nater et al. (1977) investigated the irritant effects of PG in 48-h patch tests. A group of 98 subjects, all with histories of allergic contact dermatitis, were patch tested with 100% PG. Patches containing an unspecified volume of PG were applied to the skin on the backs of each of the test subjects. After the patches were removed, sites were scored for reactions (scale = 0 to 3 + ), and skin samples were obtained by biopsy for microscopic examination. There were 11 1 + reac~ tions noted out of the 98 subjects tested. Microscopic examination of skin samples from subjects with positive reactions indicated edema and mononuclear perivas cular infiltrate in eight of the samples. The authors concluded that the reactions observed were primary irritation reactions. The irritation/sensitization potential of PG or hexylene glycol was investigated by Kinnunen and Hannuksela (1989) using several different methods. Occlusive patches containing 30% PG were applied (48 h) to the skin of 823 patients with histories of contact dermatitis. Skin reactions were scored 1-4 hand 2 days after removal of the patches. Edema and erythema were noted in 3.8% of the patients tested. Erythema alone was noted in 7.4% of the tested patients. Additional patch tests were conducted by these authors on 22 patients who had edema/erythema reactions to PG or hexylene glycol. Serial dilutions of 1, 2, 10, and 30% PG were applied using the same methods described. A positive reaction was reported in one patient tested with 1% PG; three patients tested with 30% PG had positive reactions. The effect of PG on transepidermal water loss (TEWL), skin blood flow, and nonimmunologic contact reactions (NICR) to benzoic acid was also tested by Kinnunen and Hannuksela using a spearate panel of 8 dermatitis patients and 11 normal volunteers. Doses of 20 fl-l 50% PG were applied to the backs of the 19

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

478 COSMETIC INGREDIENT REVIEW

panelists. The eight patients received five applications at the same time; the normal volunteers received repeated applications over a 7-day period. TEWL was measured with an evaporimeter after the first applications to patients and after the last applications to normal volunteers. After a 20-JJ.l application of benzoic acid to the test site, the NICR and blood flow were measured. PG significantly increased TEWL on day 3 in the patients with dermatitis and on day 8 in the normal volunteers. Cutaneous blood flow was increased in the test subjects. The NICRs in the test subjects were not significantly different from controls. PG had irritant properties in these tests. A total of 183 patients were patch tested with 38% PG. Allergic reactions were observed in 23 patients (Huriez et al., 1966; Catanzaro and Smith, 1991). Details concerning the experimental procedure were not included in the review article by Cantanzaro and Smith (1991). A total of 86 contact dermatitis patients were tested (patch tests and repeated open application tests) with PG as well as with other common ingredients of dermatological preparations. In initial patch tests, 30.0% PG in water was applied for 24 h, under occlusive patches, using Finn chambers. Reactions were scored 0.3-2 h after patch removal according to the International Contact Dermatitis Research Group grading scale (Wilkinson et al., 1970): ?+ (doubtful reaction) to + + + (extreme reaction). In this study, + (weak, non vesicular) and + + (strong, edematous or vesicular) reactions were classified as positive. Nineteen patients with allergic reactions to 30% PG in initial patch tests were subsequently patch tested (same procedure) as follows: 1% aqueous PG (5 patients), 10% aque ous PG (2 patients), and 30% aqueous PG (12 subjects); all patients had + or + + reactions. Repeated open application tests (ROATs) were also performed on these patients. In these tests, 5% PG in a cream base was applied to a 5 x 5-cm cream on the flexor aspect of the forearm twice daily for 7 days; the cream base con tained vegetable oil, cetylstearyl alcohol, parabens, and 85% water. Positive re actions to 5% PG were observed in all 5 patients who had positive reactions to 1% PG, 1 of 2 patients with positive reactions to 10% PG, and.4 of 12 patients with positive reactions to 30% PG (Hannuksela and Salo, 1986). Three of 78 patients patch tested with 10% PG had allergic reactions (Braun, 1969; Cantanzaro and Smith, 1991). In other studies, allergic reactions were ob served in 2 of 100 patients (Fisher et al., 1971; Cantanzaro and Smith, 1991) and 13 of 330 patients (Blondeel et al., 1978; Cantanzaro and Smith, 1991) patch tested with 10% PG. Details concerning the experimental procedures for these tests were not included in the review article by Cantanzaro and Smith (1991). In a study by the North American Contact Dermatitis Group (Adams et al., 1985), 29 cutaneous reactions were observed in a population of 399 patients with cosmetic-related contact dermatitis that was patch tested (Finn chambers) with 10% aqueous PG. At test concentrations of 5 and 10% PG, 15 of 1,450 patients had positive patch test reactions (Romaguera et al., 1981). Details concerning the experimental pro cedure were not stated. In a series of patch tests from 1968 to 1983, Angelini et al. (1985) reported a total of 27 positive reactions in 3,364 patients tested with 5% aqueous PG.

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 479

Test reactions to standard patch test allergens were evaluated using 1,040 sub jects (851 atopic patient~ and 189 healthy age-matched controls). The atopic pa tients (Group I, 28-41 years old; Group II, 19-27 years old) were classified in a total of eight groups as follows: Group I 1 (79 patients) and Group II 1 (97 pa tients), patients with severe atopic dermatitis who had been hospitalized; Group I 2 (149) and Group II 2 (213), patients who had been treated as outpatients; Group I 3 (59) and Group II 3 (74), patients with mild atopic dermatitis and allergic rhinitis, allergic conjunctivitis, or asthma; and Group I 4 (110) and Group II 4 (70), patients who had been examined at the hospital for allergic rhinitis, allergic con junctivitis, or asthma without dermatitis. PG (5%) was applied, using a Finn chamber, to the upper back of each patient for 2 days, and reactions were scored on day 3. Each test site was free of dermatitis at the time of patch testing and dermatitis at other sites was said to have been in an inactive phase. Of the 851 patients patch tested with 5% PG, 2.5% had allergic reactions and 3.9% had irritant/follicular reactions (Lammintausta et al., 1992). In an epicutaneous chamber test, Hannuksela et al. (1976) investigated the irritant effects of PG. Concentrations of 2.0% PG in water were placed in cham bers on the back of each of 880 test subjects. The exposure time lasted from 20 to 24 h. Mter removal of the test chambers, the exposed sites were scored at 30 min, 2, 4, and 5 days. At the concentration tested, two (0.2%) positive reactions were noted among the panelists. The authors noted that PG often caused irritant reac tions when tested under occlusive patches and that the methods employed in this test indicated a true value for allergy to the compound. The effects of PG and other compounds used as vehicles were investigated by Mendelow et al. (1985). PG increased the allergic responses in 43 patients patch tested with 50 IJ.g of 1% nickel sulfate solution. The increase in reactions noted was greater than the number of patients who had an allergic reaction to 50% PG alone. The data indicated that PG, when used as a vehicle, influenced the amount of nickel sulfate exposed to the skin, having modified the allergic reaction poten tial. In the order of largest to smallest induction of allergic responses, the vehicles were ranked: DMSO > PG > cetomacrogol cream > yellow soft paraffin. Several additional studies concerning allergy to PG have been published (Broeckx et al., 1987; Agren-Jonsson and Magnussen, 1976; Fastner, 1981; Fisher and Branaccio, 1979; Angelini and Meneghini, 1981; Oleffe et al., 1979; Wahlberg, 1984). The results of clinical provocative tests on PG are summarized in Table 12.

Case Reports

A 26-year-old woman applied minoxidil to an area of diffuse alopecia located immediately above the nape of her neck. Mter 6 weeks of application, a pruritic papular eruption was noted in the treated area and nape of the neck. Patch test reactions to the minoxidil solution and 5% aqueous PG were positive. There was no reaction to a 2% minoxidil solution in which glycerin was the solvent (Fisher, 1990). A pruritic, macular, and papular eruption was observed on the scalp (vertex

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

..... ~ ;:! ~ :::~ ~ §· :- ~ s: TABLE 12. Clinical provocative patch testing to evaluate skin irritation and sensitization potential of Propylene Glycol (PG) ~ ,Q.. Test No. of § i substance patients Procedure Results Ref. PG (2.5, 10, 866 dermatologic patients 48-h patch test (closed or cov Skin irritant (138 positive reactions); Warshaw and Herrmann ~ and 100%) ered patches) 5/23 and Ih positive patients had (1952) ~ positive reactions to 10 and 2.5% ~ PG PG (3.2, 10, 1,556 patients Patch test True allergy (4 patients) and irritation Hannuksela et al. (1975); Cat ~ 32, and (1,552 patients); results for 3 anzaro and Smith (1991) 100%) groups of 42 positive patients patch ~ tested with PG: 3.2% PG [9 posi tive ( +) reactions], 10% PG (12 + ), ~ and 32% PG (20+) tii PG (2, 10, 32, 38 patients Simultaneous 20- to 24-h Finn Allergic eczema and other types of Hannuksela and Forstrom and 100%) chamber applications; per eczema in 11 and 16 patients during (1978) ~ oral challenge dose of 2% initiation; challenge of 15 patients PG selected caused extensive exan ~ thema that cleared within 48 h 100% PG 84 patients Patch test Allergic reactions (5 patients) and Andersen and Storrs (1982) :stli skin irritation (7 patients) ~ 100%PG 98 eczema patients 48-h patch test Primary irrtation; 11 1 + reactions Nater et al. (1977) 50%PG 8 patients with allergic 5 simultaneous 20-f!.l applica PG had irritant properties; it signifi Kinnunen and Hannuksela contact dermatitis; 11 tions per patient; repeated cantly increased transepidermal (1989) normal subjects applications over 7-day pe water loss and increased cutaneous riod in normal subjects; at blood flow in both groups; NICRs end of exposure, benzoic were not significantly different acid applied to evaluate from controls nonimmunologic contact reactions (NICRs) to ben zoic acid 38%PG 183 patients Patch test 23 allergic reactions Huriez et al. (1966); Catan zaro and Smith (1991) Distributed for comment only -- do not cite or quote

PG (30% aque- 86 contact dermatitis pa- 24-h occlusive patch applica Allergic reactions in 19 patients Hannuksela and Salo (1986) ous) tients tions (Finn chambers) for 7 days PG (1, 10, and 19 patients with allergic 24-h occlusive patch applica Positive reactions: I% PG (5 subjects Hannuksela and Salo (1986) 30% in wa- reactions (from preced- tions of I, 10, and 30% PG with + or + + reactions); 10% PG ter; 5% in ing study) (Finn chambers) for 7 days; (2 with + or + + ), 30% PG (12 ;g cream) for 5% PG, repeated open with + or + +); incidence of + application test (0.1-ml ap reactions to 5% PG in preceding plications twice daily for 7 groups: 5/s (1% PG group); l/z (10% ~ days) PG group); and "'112 (30% PG group) t:s 30%PG 823 patients 48-h occlusive patch applica Erythema in 7.4% of patients; erythe Kinnunen and Hannuksela tion ma and edema in 3.8% of patients (1989) ~ PG (1, 2, 10, 22 patients (22 of patients 48-h occlusive patch applica Positive reactions: I% PG (I patient), Kinnunen and Hannuksela ~ and 30%) with erythema and tion 30% PG (3 patients) (1989) edema in preceding ~ study) 10% PG 78 patients Patch test 3 allergic reactions Braun (1969); Cantanzaro and ~ Smith (1991) g 10% PG I 00 patients Patch test 2 allergic reactions Fisher eta!. (1971); Cantan zaro and Smith (1991) ~ 10% aqueous 399 patients 48-h patch test using Finn 29 cutaneous reactions Adams et a!. (1985) PG chambers ~ 10% PG 330 patients Patch test 13 allergic reactions Blondeel et a!. (1978); Can "'tt tanzaro and Smith (1991) 5 and 10% PG I ,450 patients Patch test 15 positive reactions Romaguera eta!. (1981) g 5% aqueous 3,364 patients Patch tests from 1968 to 1983 27 positive reactions Angelini et a!. (1985) PG ::a 5% aqueous 851 atopic patients Finn chamber application for 2.5% of patients had allergic reac Lammintausta et a!. (1992) :::0 PG 2 days tions; 3.9% had irritant/follicular reactions ~ '""':>- 2% aqueous 880 patients Epicutaneous chamber test; 2 allergic reactions Hannuksela et a!. (1976) :! t:s PG chambers applied for ~ 20-24 h ~ ::::: l:li ~ ;:;· "'c ~ .- ~ ~ ..... g ;... V:l ~ ?-- ..... & ~ ...... Distributed for comment only -- do not cite or quote

482 COSMETIC INGREDIENT REVIEW region) of a rnan who had applied minoxidil to this area for 1 month. Patch test reactions to the minoxidil solution and 5% PG were positive. However, there was no reaction to 5% minoxidil in glycerin. The eruption recurred when 5% PG was rubbed into a small area of the scalp after the initial reaction subsided (Fisher, 1990). A 52-year-old woman suffered from chronic otitis media for 2 years and had used various ear drops and ointments. During the last 2 weeks of the 2-year episode, the patient used eardrops containing 2.5% hydrocortisone in a mixture of PG (50%) and water. Patch test results indicated a papulovesicular reaction (+++)to this product after 72 h. There were no reactions to 10% aqueous PG and 1 and 10% hydrocortisone in ethanol. However, oral challenge with 5 ml of PG caused a pruritic macular rash on the abdomen, a flare-up of the external ear dermatitis, and a flare-up of the patch test reaction on the back within 20 h. When the patient was patch tested 1 week after the oral challenge, a + reaction to 50% PG and a + + reaction to undiluted PG were noted after 72 h (Frosch et al., 1990). Several case studies have been published, documenting assorted reactions to PG administration in humans (Arulanantham and Genel, 1978; Demey et al., 1988; Fligner et al., 1985; Kelner and Bailey, 1985; Lolin et al., 1988; MacDonald et al., 1987; Martin and Finberg, 1970).

SUMMARY PG is an aliphatic alcohol and PPG is a polymer ofPG and water. In cosmetics, PG is used as a skin-conditioning agent-humectant, solvent, viscosity-decreasing agent, and humectant. PPGs are used as miscellaneous skin-conditioning agents. Product formulation data submitted to FDA in 1984 indicated that PG was used in a total of 5,676 cosmetic products at concentrations up to >50%. Data (FDA, 1984) submitted on PPGs were as follows: PPG 9 (6 products), PPG 26 (10 prod ucts), and PPG 425 (1 product). All three ingredients were used at concentrations up to 5%. While concentration of use data are no longer reported to FDA by the cosmetics industry, current frequency of use data reported indicate that PG is used in a total of 4,892 cosmetic products and PPG 9 in 5 products. There are no reported uses of PPG 26 and PPG 425. Current data on PG supplied to the CTFA indicate that concentrations of use range between 3 and 5% in products manufactured by one company. In mammals, the major route of PG metabolism is to lactaldehyde and then lactate via hepatic alcohol and aldehyde dehydrogenases. When PG was admin istered intravenously to human subjects (patients), elimination from the body occurred in a dose-related manner. The results of animal studies on PPGs 425, 1025, and 2025 indicate that they are readily absorbed from the GI tract and are excreted in the urine and feces. The cytotoxicity of human natural killer cells was decreased significantly in an assay in which target cells (cultured K562 erythroleukemia cells) were incubated with 1% PG. PG was relatively harmless (LD50 = 21 g/kg) in acute oral toxicity studies

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 483 involving rats. Acute oral toxicity studies on PPGs of various molecular weights (300-3,900) have indicated LD50 values (rats) ranging from 0.5 to >40 g/kg. In acute dermal toxicity studies involving groups of five albino rabbits, doses of PPG 1025 (20 ml/kg) and PPG 2025 (20 mllkg) did not cause death. Two of five rabbits dosed with 20 ml/kg PPG 425 and one of five dosed with 10 ml/kg PPG 425 died. In subchronic oral toxicity studies, PPG 2000 induced, at most, slight decreases in growth and body weight effects in rats. PPG 750 caused slight increases in liver and kidney weights in rats. Following the subchronic oral administration of PPG 750 to dogs, slight increases in liver and kidney weights were noted. In a subchronic dermal toxicity study (rabbits), PPG 2000 did not cause any adverse effects at doses of 1 ml/kg. Slight depression of growth was observed after the administration of 5- and 10-ml/kg doses. Test substance-related lesions were not observed in rats that were fed diets containing 50,000 ppm PG (2.5 g/kg/day) for 15 weeks or in rats that were fed PG concentrations up to 50,000 ppm in the diet for 2 years. Similar results were reported in a study in which dogs were fed 2 or 5 g/kg PG in the diet for -103 weeks. In another subchronic study, dogs received 5% PG in drinking water for 5-9 months. The results of tests for hepatic and renal impairment were negative. However, in cats fed diets containing PG, erythrocyte destruction was noted at concentrations as low as 6% PG. PG did not induce corneal damage in rabbits in the Draize test and was classi fied as a slight ocular irritant in another ocular irritation study. PPGs 425, 1025, and 2025 were classified as harmless agents in rabbits in another ocular irritation study; PPG 1200 induced slight, transient ocular irritation in an albino rabbit. In a 24-h skin irritation test involving nude mice, there were no reactions to 10% PG. Hypertrophy, dermal inflammation, and proliferation were observed at a concentration of 50% PG. Draize test results indicated that PG was, at most, a mild skin irritant when applied for 24 h to abraded and intact skin of rabbits. When PG was applied to the skin of guinea pigs and rabbits (guinea pigs and rabbits lack sweat glands) for 48 h using open and closed patches, no reactions were observed. The results of 48-h and 21-day open and closed patch tests involving Gottingen swine (no sweat glands) indicated no reactions to PG. Results were negative for 100% PG in a mouse external ear swelling sensitiza tion test. The results of a GPMT, OET, and chamber (Finn chamber) test indi cated no sensitization reactions to 70% PG. In another maximization test, PG was classified as a potentially weak sensitizer. The results of six other guinea pig sensitization tests indicated that PG was not an allergen. Single and repeated applications of PPG 425, PPG 1025, and PPG 2025 did not cause skin irritation in the rabbit. Repeated applications of PPG 1200 to rabbits caused mild reactions at abraded skin sites and no reactions at intact sites. PG was not teratogenic in female CD-1 mice when administered at a concen tration of 10,000 ppm on days 8-12 of gestation. Malformations were observed in 5 of226living fetuses from female mice injected subcutaneously with PG (dose = 0.1 mg/g body wt on day 9, 10, 11 of gestation). However, 3 fetuses with malfor-

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

484 COSMETIC INGREDIENT REVIEW mations were also noted among 1,026 living fetuses from the untreated control group of pregnant mice. In a continuous breeding reproduction study, there were no significant differ ences between control and experimental groups of albino mice with respect to the following: mating index, fertility index, mean number of live pups per litter, proportion of pups born alive, and sex of pups born alive. Embryonic development was reduced in cultures of mouse zygotes exposed to 3.0 M PG and inhibited completely in cultures exposed to 6.0 M PG for 20 min. In the Ames test, PG was not mutagenic in strains TA 1535, TA 1537, TA 1538, TA 98, and TA 100 of Salmonella typhimurium with and without metabolic acti vation. PG caused a dose-dependent increase in the frequency of SCEs in a Chinese hamster cell line and was classified as a weak inducer of SCEs. In another study, PG was not mutagenic when tested in an SCE assay involving human cultured fibroblasts and a cultured Chinese hamster cell line both with and without metabolic activation. Chromosomal aberrations were induced in Chinese hamster fibroblasts in another assay. PG was not mutagenic in additional in vitro tests: chromosomal aberrations, mitotic recombination, basepair substitution, micro nucleus test, reverse mutation, and DNA damage. PG disturbed the proliferation of urinary bladder epithelial cells from the rat, having reduced DNA production in tetraploid cells 1 and 2 months after the rats were injected subcutaneously. This effect was not observed at 3 months. The results were negative when PG was tested in the hamster embryo cell transformation bioassay. In a 2-year feeding study involving CD strain rats, PG was not carcinogenic when concentrations up to 50,000 ppm were administered in the diet. In a lifetime dermal carcinogenicity study, three groups of Swiss mice received dermal applications of 10, 50, and 100% PG, respectively. The tumor incidence in each of the three groups did not differ from that noted in the negative control group; skin tumors were not observed. PG induced skin irritation and sensitization reactions in normal subjects and in patients. In these studies test concentrations ranged from 2 to 100% PG. Reac tions were observed at concentrations as low as 10% PG in predictive tests and as low as 2% in provocative tests. PG also increased the allergic responses in 43 patients patch tested with 50 ILg of 1% nickel sulfate solution. Neither skin irritation nor sensitization reactions were observed in 300 subjects who received continuous and repeated dermal applications of undiluted PPG 2000.

DISCUSSION Because of the results of human irritation and sensitization tests, establishing a concentration limit for PG is considered necessary. Both provocative and predic tive test data were considered in the process of making the final determination. In provocative tests, allergic reactions were observed in 2 of 880 (0.2%) eczema patients patch tested with 2% aqueous PG, in 13 of 330 (4%) patients patch tested with 10% PG, and in 21 of 851 (2.5%) atopic patients patch tested with 5% PG. Thirty-three (3.9%) of the 851 atopic patients also had irritant/follicular reactions.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 485

Furthermore, in a study of the North American Contact Dermatitis Group, 29 cutaneous reactions were observed in a population of 399 patients with cosmetic related dermatitis who were patch tested with 10% aqueous PG. Patients with diseased skin may be at risk with respect to developing irritation/sensitization reactions to PG, even at low concentrations. Predictive tests are considered more appropriate in establishing concentration limits that are based on skin irritation or sensitization data. In these tests, normal subjects were patch tested with PG concentrations ranging from 1 to 100%. In one of the studies, skin irritation was observed when 100% PG was tested under occlusive patches, but was not observed in open patch tests. When 50 and 100% PG were tested under occlusive patches on 16 subjects, there were no reactions to 50% PG and three 1 + reactions to 100% PG. In other studies, 24 normal subjects patch tested (closed patches) with 1, 3, 10, and 30% PG had skin irritation reac tions only at concentrations of 10 and 30%, and neither skin irritation nor sensi tization was observed in 204 subjects patch tested with 12% PG (under occlusive patches) in a cream vehicle. These studies indicate that in normal subjects, PG may be a skin irritant when tested under occlusive patches and that the skin irritation potential of this ingredient may be concentration dependent.

CONCLUSION On the basis of the data included in this report, the CIR Expert Panel concludes that Propylene Glycol and Polypropylene Glycols are safe for use in cosmetic products at concentrations up to 50.0%.

Acknowledgment: The Technical Analysis was prepared by Wilbur Johnson, Jr., Senior Scientific Analyst and Writer.

REFERENCES

Abe S, Sasaki M. (1982) SCE as an index of mutagenesis and/or carcinogenesis. Prog Top Cytogenet 2:461-514. Adams RM, Maibach HI, Clendenning WE, et al. (1985) A five-year study of cosmetic reactions. JAm Acad Dermato/13:1062-9. Agren-Jonsson S, Magnussen B. (1976) Sensitization to propantheline bromide, trichlorocarbanilide, and propylene glycol in an antiperspirant. Contact Dermatitis 2:79-80. Ahluwalia P, Amma KP. (1984) Hypolipidemic effect of propane-1,2-diol on the morphology of rat erythrocytes. Res Bull Panjab Univ Sci 35:157-60. AMA Division of Drugs. (1983) AMA Drug Evaluations. 5th ed. Chicago: American Medical Associ ation, 1391. American Industrial Hygiene Association. (1980) Workplace environmental exposure level guide. Polypropylene glycols. Am Ind Hyg Assoc J 4l:A53-5. Ames BN, McCann J, Yamasaki E. (1975) Methods for detecting carcinogens and mutagens with the salmonella/mammalian-microsome mutagenicity test. Mutat Res 31:347-64. Amma KP, Namagiri T, Chum A. (1984) The redox state ofpropane-1,2-diol fed rat erythrocytes. Res Bull Panjab Univ Sci 35:109-14. Andersen, KE, Storrs FJ. (1982) Hautreizungen durch propyleneglykol. Hautarzt 33:12-4. Angelini G, Meneghini CL. (1981) Contact allergy from propylene glycol. Contact Dermatitis 7:197-8. Angelini G, Vena GA, Meneghini CL. (1985) Allergic contact dermatitis to some medicaments. Con- tact Dermatitis 12:263-9. Armitage P. (1955) Test for linear trends in proportions and frequencies. Biometrics 11:375.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

486 COSMETIC INGREDIENT REVIEW

Arulanantham K, Genel M. (1978) Central nervous system toxicity associated with ingestion of pro pylene glycol. J Pediatr 93:515-6. Bartsch W, Sponer G, Dietman K, Fuchs G. (1976) Acute toxicity of various solvents in the mouse and rat. LD50 of ethanol, diethylacetamide, dimethylformamide, dimethyl-sulfoxide, glycerine, N methylpyrrolidone, polyethylene glycol 400, 1,2-propanediol and Tween 20. Arzneimittelforsch 26:1581-3. Bauer MC, Weiss DJ, Perman V. (1992) Hematologic alterations in adult cats fed 6 or 12% propylene glycol. Am J Vet Res 53:69-72. Bekeris L, Baker C, Fenton J, et al. (1979) Propylene glycol as a cause of elevated serum osmolality. Am J Clin Pathol 72:633-6. Blondeel A, Olef'fe J, Achten G. (1978) Contact allergy in 330 dermatological patients. Contact Der- matitis 4:270-6. Bossaert L, Demey H. (1987) Propylene glycol intoxication. Arch Intern Med l47:6ll-2. Braun W (1969) Kontaktallergien gegen poylaethylenglykole. Z Haut Geschlechstkr 44:385-8. Braun AG, Buckner CA, Emerson DJ, Nichinson BB. (1982) Quantitative correspondence between the in vivo and in vitro activity of teratogenic agents. Proc Natl Acad Sci USA 79:2056--60. Brazeau GA, Fung H-L. (1989) Use of an in vitro model for the assessment of muscle damage from intramuscular injections: in vitro-in vivo correlation and predictability with mixed solvent sys tems. Pharm Res 6:766-71. Broeckx W, Blondeel A, Dooms-Goossens A, Achten G. (1987) Cosmetic intolerance. Contact Der matitis 16:189-94. Budden R, Kuehl UG, Bahlsen J. (1979) Experiments on the toxic, sedative, and muscle relaxant potency of various drug solvents in mice. Methods Toxicoll02:467-74. Catanzaro JM, Smith JG. (1991) Propylene glycol dermatitis. JAm Acad Dermatol24:90-5. Christopher MM, Perman V, Eaton JW (1989) Contribution of propylene glycol induced Heinz body formation to anemia in cats. JAm Vet Med Assoc 194:1045-56. Christopher MM, Eckfeldt JH, Eaton JW. (l990a) Propylene glycol ingestion causes o-lactic acidosis. Lab Invest 62:ll4-l8. Christopher MM, White JG, Eaton JW. (1990b) Erythrocyte pathology and mechanisms of Heinz body-mediated hemolysis in cats. Vet Pathol27:299-310. Clark CR, Marshall TC, Merickel BS, Sanchez A, Brownstein DG. (1979) Toxicological assessment of heat transfer fluids proposed for use in solar energy applications. Toxicol Appl Pharmacal 51: 529-35. Cochran WG. (1954) Some method for strengthening the common x2 test. Biometrics 10:417. Code of Federal Regulations. (April l, 1984) Title 21, Part 720.4. Voluntary filing of cosmetic product ingredient and cosmetic raw material composition statement. Information requested about cos metic products. Washington, D.C.: U.S. Government Printing Office. Code of Federal Regulations. (April I, 1984) Title 21, Parts 169.175,582.1666. Washington, D.C.: U.S. Government Printing Office. Code of Federal Regulations. (April I, 1984) Title 21, Part 175.300. Indirect food additives: adhesives coating and components. Resinous and polymeric coatings. Washington, D.C.: U.S. Government Printing Office. Commens CA. (1990) Topical propylene glycol and hyperosmolality. Br J Dermatoll22:77-80. Cosmetic, Toiletry, and Fragrance Association. (1992) Submission of unpublished data by CTFA. Concentration of use data on Propylene Glycol.* Damien M, Luciano AA, Peluso JJ. (1989) Propanediol-induced alterations in membrane integrity, metabolism and developmental potential of mouse zygotes. Hum Reprod 4:969-74. Damien M, Luciano AA, Peluso JJ. (1990) Propanediol alters intracellular pH and developmental potential of mouse zygotes independently of volume change. Hum Reprod 5:212-6. Dean ME, Stock BH. (1974) Propylene glycol as a drug solvent in the study of hepatic microsomal enzyme metabolism in the rat. Toxicol Appl Pharmacal 28:44-52. Demey H, Bossaert L. (1987) Propylene glycol intoxication and nitroglycerin therapy. Crit Care Med 15:540. Demey HE, Daelmans RA, Verpooten GA, De Broe ME. (1988) Propylene glycol-induced side effects during intravenous nitroglycerin therapy. Intensive Care Med 14:221-6. Denning DW, Webster AD. (1987) Detrimental effect of propylene glycol on natural killer cell and neutrophil function. J Pharm Pharmacal 39:236-8. Descotes J. (1988) Identification of contact allergens: the mouse ear sensitization assay. J Toxicol Cutaneous Ocul Toxicol 7:263-72. Dewhurst F, Kitchen DA, Calcutt G. (1972) The carcinogenicity of some 6-substituted benzo(a)pyrene derivatives in mice. Br J Cancer 26:506-8.

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 487

Dorman DC, Haschek WM. (1991) Fatal propylene glycol toxicosis in a horse. JAm Vet Med Assoc 198:1643-4. Dossou G, Sicard C, Kalopissis G, Raymond D, Schaefer H. (1985) Method for assessment of experimental allergy in guintsilver sulfadiazine therapy. lAMA 253:1606-9. Floersheim GL. (1985) Protection against acute ethanol toxicity in mice by zinc aspartate, glycols, and pyritinol. Agents Actions 16:580-4. Food and Drug Administration. (1984) Cosmetic product formulation data: ingredients used in each product category. Computer printout. Washington, D.C. Food and Drug Administration. (December 2, 1991) OTC drug review ingredient status report. Food and Drug Administration, Rockville, MD. Food and Drug Administration. (1992) Food additive safety profile. Section E. Safety of the food additive Polypropylene Glycol 1200. Submitted by FDA: FOI request dated 5/4/92. Food and Drug Administration. (1993) Cosmetic product formulation data. FDA computer printout. Food Chemicals Codex. (1981) 3rd ed. National Academy of Sciences, 233-4, 255-6. Fort FL, Heyman lA, Kesterson JW. (1984) Hemolysis study of aqueous polyethylene glycol 400, propylene glycol and ethanol combinations in vivo and in vitro. J Parenter Sci 38:82-7.

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

488 COSMETIC INGREDIENT REVIEW

Frosh PJ, Pekar U, Enzmann MD. (1990) Contact allergy to propylene glycol. Do we use the appro priate test concentration? Dermatol Clin 8:111-3. Fujino H, Chino T, Imai T. (1965) Experimental production of labial and lingual carcinoma by local application of 4-nitroquinoline N-oxide. JNCI 35:907-18. Gaunt IF, Carpanini FM, Grasso P, Lansdown AB. (1972) Long-term toxicity of propylene glycol in rats. Food Cosmet Toxicol10:151-62. Gebhardt CO. (1968) The teratogenic action of propylene glycol (propane-1,2-) and propanediol-1,3 in the chick embryo. Teratology 1:153-61. Glasgow AM, Boeckx RL, Miller MK, et al. (1983) Hyperosmolality in small infants due to propylene glycol. Pediatrics 72:353-5. Green S. (1977) Present and future uses of mutagenicity tests for assessment of the safety of food additives. J Environ Pathol Toxicol1:49-54. Guillot JP, Gonnet JF. (1985) The epicutaneous maximization test. Curr Probl Dermatol14:220-47. Guillot JP, Gonnet JF, Clement C, Caillard L, Truhaut T. (1982a) Evaluation of the ocular irritation potential of 56 compounds. Food Chern Toxicol20:573-82. Guillot JP, Gonnet JF, Clement C, Caillard L, Truhaut R (1982b) Evaluation of the cutaneous irritation potential of 56 compounds. Food Chern Toxicol 20:563-72. Guillot JP, Gonnet JF, Clement C, Faccini JM. (1983) Comparative study of methods chosen by the Association Francaise de Normalisation (AFNOR) for evaluating sensitizing potential in the al bino guinea pig. Food Chern Toxicol21:795-805. Halasez NA, Collins GM. (1984) Studies in cryoprotection II: propylene glycol and glycerol. Cryobi ology 21:144-7. Hannuksela M, Forstrom L. (1978) Reactions to peroral propylene glycol. Contact Dermatitis 4:41-5. Hannuksela M, Kousa M, Pirila V. (1976) Allergy to ingredients of vehicles. Contact Dermatitis 2:105-10. Hannuksela M, Pirila V, Salo OP. (1975) Skin reactions to propylene glycol. Contact Dermatitis 1:112-6. Hannuksela M, Salo H. (1986) The repeated open application test (ROAT). Contact Dermatitis 14:221-7. Hanzlik PJ, Mewman HW, Van Winkle W, Lehman AJ, Kennedy NK. (1939) Toxicity, fate and excretion of propylene glycol and some other glycols. J Pharmacal Exp Ther 67:101. Hayashi M, Kishi M, Sofuni T, lshidate M. (1988) Micronucleus tests in mice on 39 food additives and eight miscellaneous chemicals. Food Chern Toxicol26:487-500. Hickman JR. (1965) Acute toxicity of radiation sterilized propylene glycol. J Pharm Pharmacal 17:255-6. Hickman JA, Rogers QR, Morris JG. (1990) Effect of diet on Heinz body formation in kittens. Am J Vet Res 51:475-8. Hoenig V, Werner F. (1980) Is propylene glycol an inert substance? Toxicol Lett 5:389-92. Huriez A, Martin P, Vanoverschelde M. (1966) L'allergie au propylene-glycol. Bull Soc Fr Dermatol Syph 73:263-7. Informatics, Inc. (1973) Scientific literature review on generally recognized as safe (GRAS) food ingredients, propylene glycol and derivatives. Prepared for FDA under contract no. 72-104. NTIS no. PB-221 233. Ingram AJ, Grasso P. (1975) Patch testing in the rabbit using a modified human patch test method. Br J Dermatol109:58-9. Instituto Ganassini Sna di Ricerche Biochimiche. (1992) Submission of unpublished data by CTFA. UV spectral analysis of propylene glycol. (1 page).* Ishidate M, Sofuni T, Yoshikawa K, et al. (1984) Primary mutagenicity screening of food additives currently used in Japan. Food Chern Toxicol22:623-36. Jacaruso RB, Barletta MA, Carson S, Trombetta LD. (1985) Release of histamine from rat peritoneal cells in vitro as an index of irritation potential~ J Toxicol Cutan Ocul Toxicol 4:39-48. Kaldor G, Hsu Q, Wachsberger P, Chowrashi P. (1971) Effect of polyethylene glycols on contraction of various muscle models. Am J Physiol 220:639-45. Kavlock RJ, Short RD, Chernoff N. (1987) Further evaluation of an in vivo teratology screen. Tera togenesis Carcinog Mutagen 7:7-16. Kelner MJ, Bailey DN. (1985) Propylene glycol as a cause of lactic acidosis. J Anal Toxicol 9:40-2. Kero M, Hannuksela M. (1980) Guinea pig maximization test, open epicutaneous test and chamber test in induction of delayed contact hypersensitivity. Contact Dermatitis 6:341-4. Kinnunen T, Hannuksela M. (1989) Skin reactions to hexylene glycol. Contact Dermatitis 21:154-8. Klecak G, Geleik H, Frey JR. (1977) Screening of fragrance materials for allergenicity in the guinea pig. I. Comparison of four testing methods. J Soc Cosmet Chern 28:53.

JAm Coli Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 489

Kohn KW, Friedman CA, Ewig RA, lgbol ZM. (1974) DNA chain growth during replication of asynchronous L1210 cells. Alkaline elution of large DNA segments from cells lysed on filters. Biochemistry 13:4134-9. Kulick Ml, Lewis NS, Basnal, V, et al. (1980) Hyperosmolality in the bum patient: analysis of an osmolal discrepancy. J Trauma 20:223-8. Lammintausta K, Kalimo K, and Fagerlund V-L. (1992) Patch test reactions in atopic patients. Con- tact Dermatitis 26:234-40. · Lashmar UT, Hadgraft J, Thomas N. (1989) Topical application of penetration enhancers to the skin of nude mice: a histopathological study. J Pharm Pharmaco/41:118-22. Layton DW, Mallon BJ, Rosenblatt DH, Small MJ. (1987) Deriving allowable daily intakes for sys temic toxicants lacking chronic toxicity data. Regul Toxicol Pharmaco/7:96--112. Lolin Y, Francis DA, Flanagan RJ, Little P, Lascelles PT. (1988) Cerebral depression due to propylene glycol in a patient with chronic epilepsy-the value of the plasma osmolal gap in diagnosis. Postgrad Med 64:610-13. Ludvigsen CW, Thurn JR, Pierpont GL, Eckfeldt JH. (1983) Kinetic enzymic assay forD(- )-lactate, with use of a centrifugal analyzer. Clin Chern 29:1823. MacDonald MG, Getson PR, Glasgow AM, Miller MK, Johnson EL. (1987) Propylene glycol: in creased incidence of seizures in low birth weight infants. Pediatrics 79:622-5. Magnusson B, Kligman A. (1%9) The identification of con-contact allergens by animal assay. The guinea pig maximization test. J Invest Dermatol 52:268. Maguire HC. (1973) The bioassay of contact allergens in the guinea pig. J Soc Cosmet Chern 24:151. Martin G, Finberg L. (1970) Propylene glycol: a potentially toxic vehicle in liquid dosage form. J Pediatr 77:877-8. Marzulli FN, Maibach HI. (1973) Antimicrobials. Experimental contact sensitization in man. J Soc Cosmet Chern 24:399-421. Maurer T, Thomann P, Weirich EG, Hess R. (1975) The optimization test in the guinea pig. A method for the predictive evaluation of the contact allergenicity of chemicals. Agents Actions 5:174. Mendelow AY, Forsyth A, Florence AT, Baillie AJ. (1985) Patch testing for nickel allergy. The influence of the vehicle on the response rate to topical nickel sulfate. Contact Dermatitis 13:29-33. Mochida K, Gomyoda M. (1987) Toxicity of ethylene glycol, diethylene glycol and propylene glycol to human cells in culture. Bull Environ Contam Toxico/38:151-53. Mochida K, Goto M, Saito K. (1983) Effects of diphenyl, o-phenylphenol, and 2-(4'-thiazolyl) benzimidazole on growth of cultured mammalian cells. Bull Environ Contam Toxicol 31:428-31. Morrissey RE, Lamb JC, Morris RW, et al. (1989) Results and evaluations of 48 continuous breeding reproduction studies conducted in mice. Fund Appl Toxico/13:747-7. Morshed KM, L'Helgoualch A, Nagpaul JP, Amma MKP, Desjeux JF. (1991) The role of propylene glycol metabolism in lactatemia in the rabbit. Biochem Med Metab Bio/46:145-51. Morshed KM, Nagpaul JP, Majumder S, Amma MK. (1988) Kinetics of propylene glycol elimination and metabolism in the rat. Biochem Med Metab Bio/39:90-7. Motoyoshi K, Nozawa S, Yoshimura M, Matsuda K. (1984) Safety of propylene glycol and other humectants. Cosmet Toi/99:83-91. Nater JP, Baar AJ, Hoedemaeker PJ. (1977) Histological aspects of skin reactions to propylene glycol. Contact Dermatitis 3:181-5. Nelson EB, Egan JM, Abernethy DR. (1987) The effect of propylene glycol on antipyrine clearance in humans. Clin Pharmacol Ther 41:571-3. Nikitakis JM. (ed.) (1988) Cosmetic Ingredient Handbook. Washington, D.C.: Cosmetic, Toiletry, and Fragrance Association, 344, 355. Nikko Chemicals Co., Ltd. (1991-92) Cross Reference Index. An Index Cross-Referencing the English Names of Cosmetic Ingredients in Japan with CTFA Adopted Names. VI. Tokyo: Nikko Chem icals Co., 20-1. Nomura T. (1977) Similarity of the mechanism of chemical carcinogen-initiated teratogenesis and carcinogenesis in mice. Cancer Res 37:969-73. National Technical Information Service. (1979) Investigation of selected potential environmental con taminants: ethylene glycol, propylene glycols and butylene glycols. Report no. PB80-109119. Oleffe JA, Blondeel A, de Coninck A. (1979) Allergy to chlorocresol and propylene glycol in a steroid cream. Contact Dermatitis 5:53-4. Patel D, Welsh D, Baker M. (1985) Comparative study of propylene glycol and caprylic/capric tri glyceride vehicles for topical application. J Soc Cosmet Chern 36:303-11. Pegg DE, Jacobson lA, Diaper MP, Foreman J. (1987) Perfusion of rabbit kidneys with solutions containing propane-1,2-diol. Cryobiology 24:420-8.

JAm Col/ Toxicol, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

490 COSMETIC INGREDIENT REVIEW

Pienta RJ. (1980) Evaluation and relevance of the Syrian hamster embryo cell system. Appl Methods Oncol3:149-69. Quinn DA, Robinson D, Hales CA. (1990) Intravenous injection of propylene glycol causes pulmonary hypertension in sheep. J Appl Physio168:1415-20. Ramstad T, Nestrick TJ, Stehl RH. (1978) Determination of polypropylene glycol extracted from polymers into food-simulating solvents. Anal Chern 50:1325-7. Romaguera C, Garcia-Perez A, Moran M, et a!. (1981) Propylene glycol in standard patch tests. Contact Dermatitis 7:346. Rothschild L, ed. (1988) The Food Chemical News Guide to the Current Status ofFood Additives and Color Additives. Washington, D.C.: Food Chemical News, 372.2. Ruddick JA. (1972) Toxicology, metabolism, and biochemistry of 1,2-propanediol. Toxicol Appl Phar macol21:102-11. Saini M, Meenakshi, Morshed KM, Amma MKP. (1987) Propane-1,2-diol induced changes in plasma proteins and enzymes on acute oral ingestions in female rats. Res Bull Panjab Univ Sci 38:79-86. Sasaki M, Sugimura K, Yoshida MA, Abe S. (1980) Cytogenetic effects of 60 chemicals on cultured human and Chinese hamster cells. Kromosomo (Tokyo) 20:574-84. Sax NI. (1979) Dangerous Properties ofIndustrial Materials. New York: Van Nostrand Reinhold, 943. Shaffer CB, Carpenter CP, Critchfield FH, Nair JH, Franke FR. (1951) A toxicological study of some polypropylene glycols. AMA Arch Ind Hyg Occup Med 3:448-53. Shideman FE, Procita L. (1951) Some pharmacological actions of polypropylene glycols of average molecular weight 400, 750, 1200, and 2000. J Phqrmacoll03:293-305. Singh PP, Junnarkar A Y, Seshagirirao C, et al. (1982) A pharmacological study of propane-I ,2-diol. Arzneimittelforsch 32:1443-6. Speth PA, Vree TB, Neilen NF, et a!. (1987) Propylene glycol pharmacokinetics and effects after intravenous infusion in humans. Ther Drug Manit 9:255-8. Staples R, Misher A, Wardell JJ. (1967) Gastrointestinal irritant effect of glycerol as compared with sorbitol and propylene glycol in rats and dogs. J Pharm Sci 56:398-400. Stenback F, Shubik P. (1974) Lack of toxicity and carcinogenicity of some commonly used cutaneous agents. Toxicol Appl Pharmacal 30:7-13. Sweet DV. (1987) Registry of Toxic Effects of Chemical Substances. Washington, D.C.: U.S. Gov ernment Printing Office. Swenberg JA, Petzold GL, Harbach PR. (1976) In vitro DNA damage/alkaline elution assay for predicting carcinogenic potential. Biochem Biophys Res Commun 72:732-8. Thomas PA, Bhat KS, Kotian KM. (1980) Antibacterial properties of dilute formocresol and eugenol and propylene glycol. Oral Surg 49:166-70. Trancik RJ, Maibach HI. (1982) Propylene glycol: irritation or sensitization? Contact Dermatitis 8:185-9. Van Winkle W, Newman HW. (1941) Further results of continued administration of propylene glycol. Food Res 6:509-16. Wahlberg JE. (1984) Erythema-inducing effects of solvents following epicutaneous administration to man-studied by laser Doppler flowmetry. Scand J Work Environ Health 10:159-62. Wahlberg JE, Nilsson G. (1984) Skin irritancy from propylene glycol. Acta Derm Venereal (Stockh) 64:286-90. Wakabayashi T, Adachi K, Popinigis J. (1991) Effects of alkyl alcohols and related chemicals on rat liver structure and function. l. Induction of two distinct types of megamitochondria. Acta Pathol Jpn 41:405-13. Wallenius K, Lekholm U. (1973a) Oral cancer in rats induced by the water-soluble carcinogen 4-ni trochinoline N-oxide. Odont Revy 24:39-47. Wallenius K, Lekholm U. (1973b) Influence of saliva on epidermal cancer in rat induced by water or fat-soluble carcinogens. Odont Revy 24:115-26. Warshaw TG, Herrmann F. (1952) Studies of skin reactions to propylene glycol. J Invest Dermatol 19:423-30. Weast RC. (1982) CRC Handbook of Chemistry and Physics. Boca Raton: CRC Press, C-480. Wei! CS, Woodside MD, Smyth HF, Carpenter CP. (1971) Results of feeding propylene glycol in the diet to dogs for two years. Food Cosmet Toxicol9:479-82. Wilkinson DS, Fregert S, Magnusson B, Bandmann HJ, Calnan CD. (1970) Terminology of contact dermatitis. From the International Contact Dermatitis Research Group. Acta Derm Venereal (Stockh) 50:287-92. Willis CM, Stephens JM, Wilkinson JD. (1988) Experimentally-induced irritant contact dermatitis. Determination of optimum irritant concentrations. Contact Dermatitis 18:20-4.

JAm Coli Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

PROPYLENE GLYCOL AND POLYPROPYLENE GLYCOLS 491

Willis CM, Stephens CJ, Wilkinson JD. (1989) Epidermal damage induced by irritants in man: a light and electron microscopic study. J Invest Dermatol93:695-9. Willis CM, Stephens CJM, Wilkinson JD. (1990) Differential effects of structurally unrelated chemical irritants on the density and morphology of epidermal CD 1 + cells. J Invest Dermatol95:711-6. World Health Organization. (1974) Seventeenth Report of the Joint Food and Agriculture Organiza tion/WHO Expert Committee on Food Additives: Toxicological Assessment of Certain Food Ad ditives with a Review of General Principles and Specifications. Report No. 539. Yu DK, Sawchuk RJ. (1987) Pharmacokinetics of propylene glycol in the rabbit. J Pharmacokinet Biopharm 15:453-72.

*Available for review: Director, Cosmetic Ingredient Review, 1101 17th Street, N.W., Suite 310, Washington, D.C. 20036

JAm Coli Toxico/, Vol. 13, No. 6, 1994 Distributed for comment only -- do not cite or quote

International Journal of ToxicoiOC)I 31 (Supplement 2) 245S.260S @TIM Author(s) 2012 Safety Assessment of Propylene Glycol, Reprinu and permission: sagepub.comljournalsPermissions.nav Tripropylene Glycol, and PPGs as Used DOl: 10.117711091581812461381 http:Jf~ uagepub.com in Cosmetics I SAGE

1 2 2 Monice M. Fiume , Wilma F. Bergfeld , Donald V. Belsito , 2 2 2 Ronald A. Hill , Curtis D. Klaassen , Daniel Liebler , 2 2 2 James G. Marks Jr , Ronald C. Shank , Thomas J. Slaga , 2 3 Paul W. Snyder , and F. Alan Andersen

Abstract Propylene glycol is an aliphatic alcohol that functions as a skin conditioning agent, viscosity decreasing agent, solvent, and fragrance ingredient in cosmetics. Tripropylene glycol functions as a humectant, antioxidant, and emulsion stabilizer. Polypropylene glycols (PPGs), including PPG-3, PPG-7, PPG-9, PPG-12. PPG-13, PPG-15, PPG-16. PPG-17, PPG-20, PPG-26, PPG-30. PPG-33, PPG-34. PPG-51, PPG-52, and PPG-69, function primarily as skin conditioning agents, with some solvent use. The majority of the safety and toxicity information presented is for propylene glycol (PG). Propylene glycol is generally nontoxic and is noncarcinogenic. Clinical studies demonstrated an absence of dermal sensitization at use concentrations, although concerns about irritation remained. The CIR Expert Panel determined that the available information support the safety of tripropylene glycol as well as all the PPGs. The Expert Panel concluded that PG, tripropylene glycol, and PPGs 2:,3 are safe as used in cosmetic formulations when formulated to be nonirritating.

Keywords propylene glycol, tripropylene glycol, PPGs

Introduction as used in cosmetics. 1 That conclusion was confirmed in 2006.2 A safety assessment of propylene glycol (PG) and polypropy Tripropylene glycol, which has not been reviewed, is lene glycols (PPGs) was published in 1994. 1 On the basis of the included in this report. Tripropylene glycol is different from available data, the Cosmetic Ingredient Review (CIR) Expert PPG-3. The PPG-n designations all acknowledge that these Panel concluded that these ingredients were safe for use in ingredients are produced in a polymerization reaction that can cosmetic products at concentrations up to 50.0%. In that lead to some different chain length compounds, since the pro assessment, the specific PPG chain lengths were not identified, cess in not end blocked. Tripropylene glycol is an ingredient however, concentration of use data were reported for PPG-9, that contains only the "3" chain length. PPG-26, and PPG 425. Currently, the International Cosmetic Ingredient Dictionary and Handbook names PPG-3, PPG-7, PPG-9, PPG-12, PPG-13, PPG-15, PPG-16, PPG-17, PPG-20, Chemistry PPG-26, PPG-30, PPG-33, PPG-34, PPG-51, PPG-52, and PPG-69. Because new studies published after the 1994 assess Definition and Structure ment are available, that address the safety ofPG and PPGs, the Propylene glycol (CAS No. 57-55-6) is an aliphatic alcohol that Expert Panel considered these data in support of the safety of conforms generally to the formula in Figure 1. 3 Tripropylene these specific PPGs currently listed in the International Cos metic Ingredient Dictionary as well as all chain lengths that may be added in the future. 1 Cosmetic Ingredient Review Scientific Analyst/Writer This report is an update of the 1994 safety assessment and, 2 Cosmetic Ingredient Review Expert Panel Member as such, it contains information that was published after the 3 Director, Cosmetic Ingredient Review 1994 assessment was issued. Corresponding Author: Dipropy1ene glycol is not included in this report since F. Alan Andersen. Cosmetic Ingredient Review, 1101 17th Street, NW. Suite it was previously reviewed in a separate report. In 1985, 412, Washington, DC 20036, USA the Expert Panel determined that dipropylene glycol was safe Email: [email protected] Distributed for comment only -- do not cite or quote

246$ lnternotional journal of Toxicology 3 I (Supplement 2)

in cosmetics. 5 According to recent information, the USP has set safety limits ofdiethylene glycol and ethylene glycol content at a maximum of 0.1 %.6 The USP grade PG manufactured by ~OH Dow contains diethylene glycol and ethylene glycol at con HO centrations that are nondetectable (quantification limit of 0.008% wt/wt). Dow also has stated that they meet or exceed Figure I. Propylene glycol. all requirements currently found in the European Pharmaco poeia, Japanese Pharmacopoeia, and Food Chemicals Codex. Two companies submitted information regarding the con centration of propylene oxide in PPGs used to make finished products. 7 Both companies report a maximum of 10 ppm H~~~o~ propylene oxide. I 0 I OH ~ c~ Use

Figure 2. Tripropylene glycol. Cosmetic Propylene glycol is used in cosmetic formulations as a skin glycol (CAS No. 24800-44-0) is an organic compound that conditioning agent (humectant or miscellaneous), viscosity 3 conforms to the formula in Figure 2. Synonyms for PG and decreasing agent, solvent, or fragrance ingredient.3 The PPGs tripropylene glycol are listed in Table 1. function primarily as skin conditioning agents, with some func The PPGs (generic CAS No. 25322-69-4) are polymers tioning as solvents. Tripropylene glycol functions as a humec of propylene oxide that conform generally to the formula in tant, antioxidant, or emulsion stabilizer. 3 Figure 3. According to the International Cosmetic Ingredient At the time of the original safety assessment, according to Dictionary and Handbook, international nomenclature cosmetic information supplied to the Food and Drug Administration ingredient (INCI) names for the PPGs refer to the average "n" (FDA) by industry as part of the Voluntary Cosmetic Registra value corresponding to the propylene oxide chain length ofthe tion Program (VCRP), PG was used in 5676 cosmetic formula polymer; that is, PPG-3 would have an average chain length of3. tions at concentrations ranging from 0% to >50%.5 Both PPG-9 (Synonyms for PPGs are also listed in Table 1.) and PPG-26 were used in 6 and 10 cosmetic formulations, As stated above, the INCI names for cosmetic PPGs refer to respectively, at concentrations of 0.1% to 5%, and PPG 425 the chain length. However, different naming conventions are (thought to be synonymous with PPG-9) was used in 1 cosmetic used in identifying PPGs and the potential for confusion exists. formulation at a concentration range of 1% to 5%. When the official INCI name for each ingredient is used, the The frequency and concentration of use ofPG has increased. name is given as PPG, dash, and then the average number of Recent VCRP data indicate that PG is used in 9094 cosmetic units, for example, PPG-3. However, the PPGs can also be formulations (out of 34 391 total formulations reported).8 identified using the average molecular weight as part of the Polypropylene glycol (chain length not specified) is reported name; this is indicated as PPG, space, average molecular to have 45 uses. Polypropylene glycol-9 is reported to be used weight, for example, PPG 200. Table 2 gives the INCI name, in 84 cosmetic formulations, and PPG-12 is used in 3, PPG-15 molecular weight name where available, and calculated mole in 1, PPG-17 in 3, PPG-26 in 2, and PPG-30 in 5 cosmetic cular weight of the PPGs. formulations. Tripropylene glycol is used in 8 formulations. A survey of current use concentrations conducted by the Per Physical and Chemical Properties sonal Care Products Council (the Council) reported that PG is used at concentrations of0.0008% to 99%.9 Propylene glycol, The physical and chemical properties of PG, tripropylene which is used in 313 of the 580 deodorant products reported to glycol, and the PPGs are summarized in Table 3. the VCRP,8 is used at concentrations of 3% to 73%; this is the greatest leave-on concentration used.9 The highest concentra tion of use ofPG is 99%, but that use is in products that will be Method Manufacture of diluted, for example bath oils, tablets, or salts. Additionally, the Tripropylene glycol (as well as dipropylene glycol) is formed Council survey results reported that PPG-9 is used at 0.05% to by sequential addition of propylene oxide to PG.4 The products 22%, PPG-12 at 1%, PPG-17 at 1% to 2%, PPG-26 at 0.2%, and are formed simultaneously and separated by distillation. PPG-34 at 20%. Tripropylene glycol is used at concentrations up to 22%; the 22% is in an underarm deodorant. Table 4 presents current product formulation data for PG, tripropylene Impurities glycol, and the PPGs. In the original safety assessment on PG, Dow Chemical Co Propylene glycol is used in hair sprays, and its effects on the recommended that US Pharmacopoeia (USP)-grade PG be used lungs that may be induced by aerosolized products containing Distributed for comment only -- do not cite or quote

Fiume et at 2475

Table I. Synonyms

Chemical Name Synonyms/Other Technical Names

Propylene glycol 5 1,2-dihydroxypropane; 2-hydroxypropanol; methyl glycol; methylene glycol; methylethyl glycol; methylethylene glycol; monopropyl glycol; monopropylene glycol; propane-I ,2-diol; 1,2-propanediol; propane-1,2-glycol; o:-propylene glycol; 1,2-propylene glycol; propyleneglycolum (EP); trimethyl glycol Tripropylene glycol 2-(2-(2-hydroxypropoxy)propoxy)propan-l-o13 PPG-n (n =average chain length) polyoxypropylene (n)3; polypropylene glycol (n)3

Abbreviations: PPG, polypropylene glycol.

Table 2. PPG INCI Names, Molecular Weight Names, and Calculated Molecular Weightsa

PPG INCI name Molecular Weight OH (PPG-n; n =avg Name as Indicated by Calculated number of moles of the Trade Name Listed Molecular Weight OLi propylene oxide) in the Dictionary (n x 58) + 18 CH3 n PPG-3 PPG 200 192 PPG-7 424 Figure 3. Polypropylene glycol. PPG-9 PPG 400 or PPG 425 540 n indicates average propylene oxide chain length and is reflected in the name, PPG-12 714 for example, PPG-12 would have n == II PPG-13 772 PPG-15 888 this ingredient may be of concern. The aerosol properties that PPG·I6 PPG 950 946 determine deposition in the respiratory system are particle size PPG-17 PPG 1000 1004 and density. The parameter most closely associated with PPG-20 PPG 1200 1178 deposition is the aerodynamic diameter, da, defined as the PPG-26 PPG 2000 1526 PPG-30 PPG 4000 1758 diameter of a sphere of unit density possessing the same ter PPG-33 1932 minal settling velocity as the particle in question. In humans, PPG-34 1990 particles with an aerodynamic diameter of $10 ).lm are respir PPG-51 2976 able. Particles with a da from 0.1 to 10 11m settle in the upper PPG-52 PPG 3000 3034 respiratory tract and particles with a da < 0.1 ).lm settle in the PPG-69 4020 10 11 lower respiratory tract. • Particle diameters of 60 to 80 11m Abbreviations: PPG, polypropylene glycol; INCl. international nomenclacure and 2:80 ).lm have been reported for anhydrous hair sprays and cosmetic ingredient, pump hairsprays, respectively.12 In practice, aerosols In original report, but not specifically listed in table: should have at least 99% of their particle diameters in the I 0 PPG 125 PPG300 to 110 ).lm range and the mean particle diameter in a typical PPG 750 13 aerosol spray has been reported as -38 j.lm. Therefore, most PPG 1025 aerosol particles are deposited in the nasopharyngeal region PPG 2025 and are not respirable. PPG 3900 Tripropylene glycol, PG, and PPGs are not included in the list of ingredients that are prohibited for use in the European Propylene glycol is used as an inactive ingredient in anum Union 14 or on the list of ingredients restricted or prohibited for ber of FDA-approved drug products. It has been approved at use in Japan. 15 concentrations up to 98.09% in topical drugs and 92% in oral solutions. 19 There is inadequate evidence to establish PG as GRAS and effective in OTC pediculicide drug products. Noncosmetic Propylene glycol has many uses in pharmaceuticals, food, Propylene glycol is generally recognized as safe (GRAS) as a and manufacturing.20 It is used in organic synthesis, especially direct food additive when used in accordance with good man for PPG and polyester resins. 21 ufacturing practices, and it is approved as a direct and indirect Polypropylene glycol is approved as a secondary direct food food additive.16 According to the Joint FAO/WHO Expert and additive and as an indirect food additive.16 Polypropylene Committee on Food Additives (JECFA), the acceptable daily glycol has many industrial uses.21 intake (ADI) ofPG is 25 mglkglbw/d. 17 In Japan, the Ministry Tripropylene glycol also has many uses in pharmaceuticals, ofHealth, Labour, and Welfare (MHLW) specified that accord food, and manufacturing. It is used as an intermediate in resins, ing to the food sanitation law, PG has no potential to cause plasticizers, pharmaceuticals, insecticides, and the production harm to human health. 18 of ethers and esters.22 Distributed for comment only -- do not cite or quote

248S International journal ofToxicology 31 (Supplement 2)

Table J. Chemical and Physical Properties

Characteristic Description

Propylene glycols Color and form Colorless viscous stable hygroscopic liquid Odor Practically odorless Molecular weight 76.0920 Solubility Miscible in water, acetone, and chloroform; soluble in ether; miscible with water, alcohol, and many organic solvents Melting point -59°C; -60°C Boiling point 187.3°C; 188.2°C Freezing point -59°C Density/specific gravity I.036 @ 25°C WC; 1.038 I @ 20°C/200C Disassociation constant (pKa) 14.8@ 25°C Octanol/water partition coefficient log Kow = -0.92 Index of refraction 1.4323 @ 20°C; I .4293 @ 27"C Tripropylene glycol 22 6 22 Color and form Colorless liquid ·s ; slightly viscous Odor Odorless22·56 Molecular weight 192.2622 Solubility Soluble in water, methanol, and ether: miscible with alcohol22; miscible in waters' Melting point <-300C22.S6 2 7 Boiling point 271°C 2.S Density/specific gravity 1.0 19 @ 20°C/20°C22 Octanollwater partition coefficient log Pow= -0.5S6'57 Index of refraction 1.4449 @ 20°C/D22 21 reactivity Combustible Polypropylene glycols Color and form Clear, colorless or practically colorless, viscous liquidsa Molecular weight Dependent on chain length Solubility Lower mol wt polymers are soluble in waters; soluble in such organic solvents as aliphatic ketones and alcohols but is insoluble in ether and most aliphatic hydrocarbons (mol wts not defined)sa pH Between 6 and 959 Density 1.002-I .00760 Reactivity Nonvolatile; combustibles

T oxicokinetics was seen at 32 minutes. The authors suggested that PG mole cules diffuse into stratum corneum only to a depth of6 to 7 )1m, Absorption approximately. The researchers also suggested that PG mole Propylene glycol. The dermal penetration of [14C)PG through cules do not reach the dermis. excised female hairless mouse skin from the ternary cosolvent containing I 0 mol% oleic acid and 6 mol% dimethyl isosorbide Dermal Penetration Enhancement in 84% PG was determined.23 Over a 24-hour period, the cumu lative penetration ofPG was 57.1 % of the applied amount. Propylene glycol. Propylene glycol has been described as a The dermal absorption of PG was determined in the outer penetration enhancer, and penetration enhancers act by various most layers of skin using thermal emission decay-Fourier mechanisms to perturb diffusional pathways through the skin. transform infrared (TED-FTIR) spectroscopy.24 Propylene gly Proposed mechanisms of penetration enhancement by PG col was applied to the fingertip of one human participant for 30 include alteration of barrier function by its effects on a keratin structure or a PG-induced increase in the solution capacity minutes using PG-soaked cotton wool. The site was wiped and 23 allowed to dry for 1 minute. The thickness of the surface layer within the stratum corneum. Examples of the effect of PG of stratum corneum probed was 0.71 J.lm. Measurements were on penetration are summarized in Table 5. performed every 25 minutes over a 3-hour period, with I mea surement taking 15 minutes. The concentration of PG remain Toxicology ing at the surface of the stratum corneum decreased over time. At 12 and 32 minutes, the maximum concentration ofPG was Cytotoxicity found at a depth of< 1 J.lm, while at 107 and 157 minutes, the Propylene glycol. The cytotoxicity of PG was determined in maximum concentration of PG was found at a depth of 3 to 4 assays that measured inhibition of human foreskin fibroblasts J.lm. At a depth of 6 J.lm, the greatest concentration ofPG, 0.2%, and keratinocytes, inhibition of collagen contraction by Distributed for comment only -- do not cite or quote

Fiume et ol 249S

Table 4. Frequency and Concentration of Use Table 4. (continued)

Freq of Cone of Freq of Cone of 8 9 8 9 Product Category Use 2009 ·" Use (%) 2009 Product Category Use 2009 ·" Use (%) 2009

Propylene glycol Nail polish and enamel 2 (24) 0.0008-6 Baby products removers Shampoos 6 (56) 0.005-0.4 Others IS (138) Not reported Lotions/oils/powders/ 18 (137) 0.02 Oral hygiene products creams Dentrifrices 4 (59) 0.02-10 Others 26 (143) 0.00 1-0.003b Mouthwashes and breath 9 (74) 0.04-5 Bath preparations fresheners Oils, tablets, and salts 23 (314) 1-99 Others 4 (86) Not reported Bubble baths 24 (169) 1-5 Personal cleanliness products Capsules I (4) Not reported Bath soaps and detergents 502 (1665) 0.01-25 Others 64 (234) Not reported Underarm deodorants 313 (580) 3-73 Eye makeup preparations Douches 4 (14) I Eyebrow pencils 3 (144) 2- 14 Feminine deodorants 9 (19) Not reported Eyeliners 94 (754) 0.2-16 Others 272 (792) 2-IOd Eye shadows 40 (1215) 0.03-18 Shaving preparations Lotions 66 (254) 0.02-47 Aftershave lotions 174 0.02-8 Makeup removers 21 (128) 0.03-2 Preshave lotions I (22) Not reported Mascaras 130 (499) 0.3-16 Shaving creams 37 (122) 4-40 Others liS (365) r Shaving soaps 3 (10) Not reported Fragrance preparations Others 59 (134) Not reported Colognes and toilet waters 304 (1377) 0.3-6 Skin care preparations Perfumes 117 (666) 0.03-5 Cleansers 398 (1446) 0.5-39 Powders 3 (221) 0.005-1 Depilatories 14 (42) 0.006-13 Others 120 (566) 0.2-70 Face and neck 558 (1583) 5-30 Noncoloring hair preparations Body and hand 648 (1744) 0.009-68 Conditioners 446 (1226) 0.08-42 Face and neck sprays No category 6 Sprays (aerosol fiXatives) 60 (312) 0.003-4 Body and hand sprays No category 1-10 Straighteners 129(178) 4-25 Foot powders and sprays II (47) 0.03 Permanent waves 7 (69) 0.3-10 Moisturizing products 846 (2508) 0.2-41 Rinses (noncoloring) 9 (33) 0.5-10 Night preparations 121 (353) 0.004-20 Shampoos 494 (1361) 0.06-5 Paste masks (mud packs) 136 (441) 0.1-11 Tonics, dressings, etc 468 (1205) 0.3-40 Skin fresheners 84 (259) 0.002-7 Wave sets II (51) Not reported Others 41S (1308) 2-20. Others 318 (807) 0.3-38 Suntan preparations Hair coloring preparations Suntan gels/creams/liquids 43 (107) 0.01-S Hair dyes and colors 1361 (2393) S-IS Indoor tanning preparations 86 (240) 1-33 Hair tints 20 (21) 10 Others 19 (62) 10 Hair rinses NR I Total for propylene glycol 9747 (34,391) 0.0008-99 Hair shampoos 16 (40) Not reported Tripropylene glycol Hair color sprays (aerosol) 7 (7) Not reported Fragrance preparations Hair lighteners 5 (21) Not reported Perfumes I (666) Not reported Hair bleaches 13 ( 149) Not reported Personal cleanliness products Others 23 (168) 6-16 Underarm deodorants 7 (S80) 21-22 Makeup preparations Skin care preparations Blushers 17 (43"1) 0.2-67 Moisturizing creams/lotions/ Not reported 0.00004 Face powders IS (661) 0.009-0.2 powders Foundations 134 (S89) 4-S7 Total for tripropylene glycol 8 0.00004-22 Leg and body paints 4 (29) 0.03-0.4 Polypropylene glycol Lipsticks 39 (1883) 0.1-8 Fragrance preparations Makeup bases 42 (117) 0.1-21 Colognes and toilet waters 30(1377) Not reported Makeup fixatives 3 (45) Not reported Perfumes 4 (666) Not reported Others 7S (48S) 2-19 Hair coloring preparations Manicuring preparations Hair dyes and colors 6 (2393) Not reported Basecoats and undercoats 3 (79) Not reported (requiring caution stmt) Cuticle softeners II (27) 4 Hair bleaches I (149) Not reported Nail creams and lotions 6 (14) 0.02-12 Makeup preparations (not eye) Nail polish and enamel 8 (333) 0.008-0.9 Blushers (all types) I (434) Not reported (continued) (continued) Distributed for comment only -- do not cite or quote

250S International journal of Toxicology 31 (Supplement 2)

Table 4. (continued) Table 4. (continued)

Freq of Cone of Freq of Cone of 9 8 Product Category Use 2009".. Use (%) 2009 Product Category Use 2009 .. Use (%) 20099

Personal cleanliness preparations Total for PPG-26 2 0.2 Others I (792) Not reported PPG-30 Shaving preparations Noncoloring hair preparations Aftershave lotions I (367) Not reported Tonics, dressings, other hair I (1205) Not reported Skin care preparations grooming aids Cleansers I (1446) Not reported Skin care preparations Face and neck preparations I (1583) Not reported Cleansers 3 (1446) Not reported (excl. shaving) Face and neck (excl. shaving) I (I 583) Not reported Suntan preparations Totals for PPG-30 s Not reported Indoor tanners I (240) Not reported PPG-34 Total for polypropylene glycol 47 Not reported Skin care preparations PPG-9 Paste masks (mud packs) Not reported 20 Bath preparation Total for PPG-34 Not reported 20 Others 3 (234) Not reported Eye makeup preparations • Total number in category given in parentheses. Eye lotions Not reported II b 0.003% in a rinse-off product. < 7% in a brow and lash gel. Noncoloring hair preparations 4 2% in a shower gel; 6% in a foot scrub. Shampoos (noncoloring) 74 (1361) 0.5 • 6% in a vaginal area moisturizer/lubricant. Personal cleanliness products Bath soaps and detergents Not reported 22 Others 33 (792) Not reported Skin care preparations fibroblasts, and changes in cell mollJhology of fibroblasts and Cleansers Not reported 0.05-0.4 keratinocytes.25 Fibroblast and keratinocyte proliferation was Depilatories Not reported 4 inhibited within 3 days after administration of PG; no signifi- Face and neck creams, Not reported IS cant changes in cell proliferation occurred with a 6-day admin- lotions, and powders istration. Propylene glycol was a moderately potent inhibitor, Skin fresheners Not reported 4 Total for PPG-9 110 0.05-22 with an IC50 (concentration causing 50% proliferation inhibi- PPG-12 tion) of 280 mmoi/L for fibroblasts and 85 mmoi/L for kerati- Noncoloring hair preparations nocytes. The effect of PG on collagen contraction by Hair conditioners 2 (1226) Not reported fibroblasts was concentration dependent throughout the entire Tonics, dressings, other hair I (1205) Not reported study. The concentration causing 50% contraction inhibition grooming aids was 180 mmoi/L. Skin care preparations The effect of PG on changes in cell morphology also was Face and neck creams, Not reported 25 lotions, and powders examined. A gradual detachment of cells from the culture Total for PPG-12 3 accompanied by changes in cell shape occurred in confluent PPG-15 keratinocyte cultures when the concentration of PG was Eye makeup preparations increased above 5%. After 24 hours, replacing medium Eyeliners I (754) Not reported containing 5% PG with PG-free medium resulted in almost Total for PPG-1 5 I Not reported complete recovery within 48 hours. However, this recovery did PPG-17 not occur with 7% PG. Similar results were observed with Skin care preparations Face and neck (excl. shaving) 3 (IS83) fibroblasts, and the concentration inducing irreversible cell Moisturizing creams, lotions, Not reported damage in both fibroblast and keratinocytes cultures was 660 and powders mmoi/L PG. Suntan preparations Suntan gels, creams, and liquids Not reported 2 Total for PPG-17 3 1-2 Single-Dose Toxicity PPG-26 Fragrance preparations Oral Perfumes I (666) Not reported Skin care preparations Polypropylene glycols. The acute toxicity of PPG 425 was Face and neck creams, Not reported 0.2 evaluated using 2 groups of 3 rats (strain and gender not lotions, and powders specified)?6 The rats were given a single oral dose of 250 or Paste masks (mud packs) Not reported 0.2 1000 mglkg PPG 425 by gavage and observed for 14 days. Others I (1308) Not reported Animals of the low-dose groups had convulsions and loss of (continued) coordination, whereas animals of the high-dose group had Distributed for comment only -- do not cite or quote

Table 5. Penetration Enhancement by PG

Test Chemical Barrier Methods Results

Dihydroergotamine (DHE) Excised male New Zealand white -In vitro diffusion using Franz diffusion cells; -the avg amount of DHE that penetrated from a water base during mesylate rabbit skin several vehicles were used to examine their 24 h was 0.3 J,lg. with a max. rate of absorption of 0.02 Jlg/h during effect on the penetration of 16.0 mg/ml DHE; the 9- to 12- hour period; with a PG-vehicle, the avg amount of the penetratioo of various coocentrations of DHE that penetrated was 7.25 J,lg. with a max rate of absorption of DHE in PG was also examined 0.3 Jlg/h during the 12- to 24-hour time period; for comparison, the amount absorbed with PEG-400 and liquid paraffin vehicles was 3.05 and 4.14 J,lg DHE over 24 hours; -the avg amount of DHE that penetrated with 8, 16, 30, and 50 mg/ml DHEin PG was 3.78, 7.25, 14.47, and 38.98 J.lg over 24 hours61 Betamethasone 17 -valerate Excised human abdominal skin -The multilayer membrane system (MMS) was -with BMV gels, PG acted as a cosolvent as penetration was (BMV); hydrocortisone 17- used to evaluate the penetration of the GCIPG thermodynamically controlled; I 0%-80% PG was evaluated-BMV butyrate HCB); hydrocorti gel formulations; time intervals for the penetration was almost unaffected with 10%-40% PG, was slightly sone (HC) (topical glucocorti penetration studies were 200 min and 24 h; increased with 40%-60% PG, and was decreased with >60% PG; the coids [GCs]) -the relationship between the physicochemical greatest amount of BMV in whole skin was found with 40% PG; properties of the drugs in binary PG/water -penetration of HCB with I0%-80% PG was evaluated, and the mixtures and the rate and extent of their greatest penetration was with 20% PG; penetration was studied; -with HC. the rate and penetration increased with increasing PG -in vitro penetration through skin was contents from 5% to 80% in suspension-type gel formulation; the determined enhancement effect of PG masked the thermodynamkally controlled behavior of HC; increased penetration of HC with increasing PG concentration was detected in the stratum comeum and the viable epidermis, but the amount permeating into the dermis was independent of PG content62 Pyrene butyric acid (PBA) Excised female human breast skin -Dermal penetration through skin was -high PBA release rates from PG were seen in the MMS; determined using glass diffusion cells; -PBA and PG appear to penetrate simultaneously into the stratum -the MMS was used for liberation studies; comeum; at I000 minutes, the ratio is increased indicating an -the effect of fatty acids was also evaluated accumulation or deposition of PBA due to more rapid PG-transfer into the epidermis; -PG penetrated better with the addition of fatty acids; mode of enhancement is non-specific37 Aspirin Excised porcine ear epidermis -in vitro studies using Franz cells, FTRI. and in all of the studies, the percutaneous absorption of aspirin was TEWL were used to determine penetration increased greatest with 80% ethanol/20% PG; these results were enhancement by PG, ethanol, and very similar to ethanol alone; PG alone did not significantly combinations of the 2 increase absorption of aspirin63 Diclofenac sodium (DfS) Excised male Wistar rat skin -DFS gels were dissolved in water, a 20%-60% -PG acted as a cosolvent. not a penetration enhancer; PG/water mixture, or 30%-40% PG/3%-5% -PG/IPM synergistically enhanced drug flux; maximum isopropyl myristate (IPM)/water (pH 7.2) enhancement ratios were with 40% PG 64

~ Distributed for comment only -- do not cite or quote

2525 International journal of Toxicology 31 (Supplement 2) convulsions. One high-dose animal died on day 1. All low-dose male dose groups, while mid- and high-dose females had sig animals and the remaining 2 high-dose animals survived until nificantly decreased body weights starting on days 64 and 50 of study termination. the study, respectively. Feed consumption was decreased for the females starting on days 50 and 43, respectively. Relevant Parenteral differences occurred in some hematological parameters, serum enzyme activities, and lung, spleen, liver, and kidney weights; Propylene glycol. An acute study was performed in which however these differences were inconsistent and without dose female ICR mice were dosed intraperitoneally (ip) with 2600, response trends. The mid- and high-dose animals had increased 7 5200, or 10 400 mglkg PG.2 All except the high-dose mice goblet cells and increased mucin within these cells. survived 6 days after dosing. (The number of high-dose mice that died was not given.) Signs of toxicity, such as lethargy and ruffled hair coats, were not observed in the 2600 and 5200 Ocular Irritation groups. Propylene glycol. The ocular irritation potential of PG was determined using groups of 6 male and female New Zealand 32 Repeated-Dose Toxicity white albino rabbits. First, a single application of 1 drop of PG was instilled into the conjunctival sac of the left eye of each Oral rabbit, and the eye was not rinsed. In the second part of the Propylene glycol. Groups of 6 inbred male Wistar rats were study, 1 drop ofPG was instilled into the conjunctival sac of the dosed orally by gavage, daily, with 294.23 mg PG/100 g body left eye every 24 hours for 3 consecutive days. At both times, wt (as 1 mL 28.4o/o/100 g) for 10 (group 1), 20 (group 2), or 30 the contralateral eye was untreated and served as the control. days (group 3), and the effects on a number of intestinal para The eyes were examined on days 1, 2, 3, and 7. With the single meters were determined.28 Control groups received an equal application, slight-to-moderate conjunctival hyperemia was volume of saline for 10, 20, or 30 days. After termination of observed on day 1 and resolved by day 2. The highest total dosing, animals were fasted overnight and then killed. All ani score was 19 of 550, well below the category of marginal mals survived until study termination. Body weight gains were irritant (score of 65). Multiple instillations resulted in similar statistically significantly decreased for animals in group 1 and observations, with slight hyperemia lasting up to day 3 in 2 increased for animals in groups 2 and 3. A number of enzyme rabbits. The highest total score following multiple installations activities were enhanced; statistically significant increases was 38 of 550, again below the category of marginal irritant. were seen in sucrase activity in groups 1 and 2 and lactase and y-glutamyl transpeptidase activity in group 3. Absorptive func Dermal Irritation/Sensitization tion was assessed by measuring nutrient uptake. Statistically significant increases of o-glucose and calcium uptake were Propylene glycol. The dermal irritation potential of 100% PG 33 seen in all groups and of glycine, L-asparate, and L-lysine was evaluated with male hairless SK.Hl hr/hr mice. Propy uptake was seen in groups I and 2. Scanning electron micro lene glycol was instilled in polyvinyl chloride cups (vol 0.3 3 scopy revealed that PG did not affect the intestinal mucosal cm ) on the dorsal side of3 mice. The test substance remained surface. in contact with the skin for 24 hours. At the end of the 24 hours, Nineteen male Han:Wistar rats were given drinking water the animals were killed and a sample of the exposed skin was containing 40 giL PG for 2 weeks; a control group of 16 rats examined microscopically. Propylene glycol was minimally was given tap water.29 The animals were placed in metabolism irritating, with a total score of 7 (maximum score = 77). cages during the last 24 hours of dosing and urine was col lected. Propylene glycol administration did not have any effect Reproductive and Developmental Toxicity on urinary excretion of oxalic or alkoxyacetic acid, nor did it affect pH or urinary metabolites. Propylene glycol did not Propylene glycol. The reproductive and developmental effects 34 cause any renal effects. ofPG were evaluated using mice, rats, rabbits, and hamsters. Groups of 8 male and 8 female CD-1 mice were given 0.5%, Groups of 25 or 28 female albino CD-1 outbred mice were 1.0%, 2.5%, 5.0%, and 10.0% PG in the drinking water for 14 mated and 22, 22, 22, 20, and 23 gravid mice were dosed by days.30 Negative controls were given untreated drinking water. oral intubation with 0.0, 16.0, 74.3, 345.0, and 1600.0 mg/kg aq Body weight gains of test animals were similar to or greater PG on days 6 to 15 of gestation. Groups of 25 to 28 female than controls. No animals died during the study. albino Wistar rats were mated and 22, 23, 22, 20, and 24 were dosed as above, respectively. Positive control groups of 23 mice and 21 rats were given 150.0 or 250.0 mglkg aspirin, Inhalation respectively. Body weights were recorded at various intervals Male and female Sprague-Dawley rats (number per group not and general observations were made daily. Caesarian sections given) were exposed to 0.16, 1.0, or 2.2 mg PG/L air for 6 were performed on days 17 and 20 for all mice and rats, respec hours/d, 5 days/week, for 13 weeks in a nose-only inhalation tively. All fetuses were examined macroscopically for visceral study. 31 There was no difference in body weights for any ofthe or skeletal defects. Administration of PG did not affect Distributed for comment only -- do not cite or quote

Fiume eta/ 253S maternal or fetal survival in mice or rats, and there were no 3 hours after hCG administration. The males were removed 16 statistically significant differences in fetal anomalies between hours after dosing with PG. Mated females were given colchine test and negative control groups in mice or rats. 22 hours after dosing with PG; zygotes were collected 18 hours Groups of 11, II, 12, 14, and 13 gravid female Dutch-belted later. There were 30, 40, 49, and 66 mice in the control, 1300, rabbits were dosed by oral intubation with 0, 12.3, 57 . I, 2600, and 5200 mglkg groups, respectively. The increase in 267.0, or 1230.0 mglkg aq PG on days 6 to 18 of gestation, hyperploidy was statistically significant in all test groups com respectively. A positive control group of 10 gravid rabbits was pared to controls. A statistically significant change was not given 2.5 mg/kg 6-aminonicotinamide. Body weights were seen for polyploidy or hypoploidy, and zygotes containing recorded at various intervals and general observations were PCS, premature anaphase, or single chromatids were not found. made daily. Caesarian sections were performed on day 29. All The authors noted that there was no statistically significant fetuses were examined macroscopically and kept for 24 hours difference in the proportion of zygotes collected for each group to evaluate survival. The pups were then examined viscerally compared to oocytes. However, the number of zygotes ana and for skeletal defects. Administration of PG did not affect lyzed compared to the number placed on slides was signifi maternal or fetal survival, and there were no statistically cantly decreased in the test groups; a relatively large portion significant differences in fetal anomalies between test and of these zygotes had clumped chromosomes. negative control group. Groups of 24-27 female golden hamsters were mated and 21, 24, 25, 22, and 22 gravid hamsters were dosed by oral Genotoxicity intubation with 0.0, 15.5, 72.0, 334.5, and 1550.0 mglkg aq Tripropylene glycol. In a preincubation study with tripropylene PG on days 6 to 10 of gestation, respectively. Positive controls glycol using Salmonella typhimurium strains TAI535, TAIOO, were given 250.0 mglkg aspirin. Body weights were recorded TA97, and TA98, the results were negative using concentra at various intervals and general observations were made daily. tions of 0 to I 0 000 J..lg/plate with and without metabolic Caesarian sections were performed on day 14. All fetuses were activation.36 examined macroscopically and for visceral or skeletal defects. Administration ofPG did not affect maternal or fetal survival, and there were no statistically significant differences in fetal Clinical Assessment of Safety anomalies between test and negative control groups. Propylene glycol. Propylene glycol dermal penetration was Propylene glycol was used as a vehicle in a reproductive and reportedly enhanced by the addition of fatty acids, such as oleic behavioral development study.35 It was administered to 15 acid. 37 Transepidermal water loss (TEWL) and attenuated total gravid Sprague-Dawley rats orally by gavage on days 7 to 18 reflectance (A TR)-FTIR were used to evaluate participants of gestation at a volume of 2 mL/kg. Propylene glycol did not exposed to PG and/or oleic acid.38 The TEWL was determined have any effects on reproductive or behavioral development using I 0 participants (number of males and females not spec parameters. ified) with application of occlusive chambers containing noth Female ICR mice were used to determine whether PG ing, 300 IlL PG, or 300 J..lL 0.16 moi/L oleic acid in PG, for 3 or induced cytogenetic aberrations in mouse metaphase 11 (Mil) 24 hours. The fourth site was not treated and not occluded. The oocytes that predispose zygotes to aneuploidy.27 Groups of TEWL measurements were started 3 hours after chamber mice were first given an ip injection of7.5 IU eCG to augment removal to reduce volatile solvents on the skin surface in order follicular maturation followed 48 hours later with 5 IU human to avoid interference with the Evaporimeter. The site treated chorionic gonadotropin (hCG) to induce ovulation. After 3 with oleic acid/PG increased water loss for a longer period in hours, mice were dosed ip with 1300, 2600, or 5200 mg/kg comparison to the PG only or empty sites. The 3- and 24-hour PG in distilled water. A control group was given distilled water applications of PG resulted in an enhanced water loss ratio of only. For the Mil portion of the study, ovulated oocytes were 1.1. With oleic acid/PG, these values were 2.0 and 2.1, collected from 20 test animals/group and 30 control animals respectively. and processed for cytogenetic analysis 16 hours after adminis For the ATR-FTIR portion, an occlusion system containing tration of PG. The number of oocytes collected from test ani PG or oleic acid/PG was applied to the forearm of each parti mals was nonstatistically significantly increased compared to cipant; a third site was untreated. The chambers were removed controls. A statistically significant change in hyperploidy, after 3 hours, and ATR-FTlR spectra were recorded. Upon hypoploidy, or single chromatids was not observed. An removal at the site where oleic acid/PG was applied, the increase in the frequency of PCS at each dose was statistically absorbance at the wavelength measuring free acid indicated significant, and the incidence of premature anaphase was sta the presence of extra free acid, while the absorbance at the tistically significantly greater in the 5200 mglkg dose group as wavelength characteristic of esterified ester lipids was similar compared to controls. Neither metaphase I nor diploid oocytes to untreated and PG-treated sites. The absorbance ratio for were found. these 2 wavelengths leveled off to that of the untreated site For the zygote portion of the study, the female mice were 3 hours after removal of the chambers, indicating migration of paired with undosed males immediately after being given hCG; oleic acid into lower cell layers or lateral spreading within the the females were dosed ip with 1300, 2600, or 5200 mg/kg PG stratum corneum. The researchers also examined ATR-FTIR Distributed for comment only -- do not cite or quote

254S International journol of Toxicology 3 I(Supplement 2)

T able 6. Clinical Dermal Irritation/Sensitization Studies With Propylene Glycol-Predictive

Dose Participants Procedure Results

Irritation 69.15% in a deodorant 20 24-h 510PT Pll-0.25; significandy less irritating than the reference control'40 68.06 in a deodorant 20 SIOPT Pll-0.1341 PG 12 occlusive chambers for 3 or 24 h; LDV irritation index-1.1 (3 h); 1.2 (24 h) -slight erythema42 0.16 moi/L oleic acid/PG 12 occlusive chambers for 3 or 24 h; LOY irritation index-2.1 (3 h); 3.9 (24 h) -clearly visible erythema42 35% in a deodorant 26M 30-day use study no potential for eliciting irritation or sensitization43 65.2% in a deodorant 40 F 30-day use study no potential for eliciting irritation or sensitization44 73% in a deodorant 24M 30-day use study no potential for eliciting irritation or sensitization65 65.8% in a deodorant 26M 4-week use study no potential for eliciting irritation or sensitization46 Sensitization 69.15% 18M; 7F maximization test no sensitization reactions 47 73% in a deodorant 30M; 71F RIPT scores of+ to 2 observed throughout the study; 4 participants discontinued durin! induction due to a repeat moderate reaction 8 86% in a deodorant 99 RIPT one + score observed throughout the study; no sensitization49

when the oleic acid!PG site was tape stripped 5 times, remov the test formulation, 4 participants had a score of± (minimal ing 50% ofthe thickness of the stratum corneum, 2 hours after faint uniform or spotty erythema) and 3 participants had a removal of the application chambers. The results indicated score of 1 (pink-red erythema visibly uniform in the entire oleic acid accumulates in a deeper layer after the tape contact area.) The primary irritation index (PII) for the deo stripping. dorant containing 69.15% PG was 0.25. This product was significantly less irritating than the reference control, which had a PH of0.93 and 17 of20 participants with scores between Derma/Irritation/Sensitization ±and 3. Propylene glycol. Intradermal injection of 0.02 mL undiluted In another SIOPT, a deodorant formulation containing PG produced a wheal-and-flare reaction within minutes, while 68.06% PG was tested undiluted using 20 participants (gender 1 the same volume applied epidermally did not produce any not specifiedt A deodorant currently in use was used as a reaction.39 These authors reported that, occasionally, subjec reference control. Three participants had a score of ± and I tive or sensory irritation sometimes occurred in volunteers after had a score of 1 to the test formulation. The PII for the test application of various concentrations ofPG. Some researchers formulation was 0.13, which was not significantly different have proposed classifying skin reactions to PG into 4 groups: than the PII of 0.15 for the reference control. (1) irritant contact dermatitis; (2) allergic contact dermatitis; The irritation index for PG and 0.16 mol/L oleic acid/PG (3) nonimmunologic contact urticaria; and (4) subjective or was determined using 12 participants (number per gender sensory irritation. not specified) by applying occlusive chambers containing these 2 test substance to the volar forearm for 3 or 24 hours.42 An empty chamber was applied to a third site, and Predictive Testing-lrritation the fourth site was an untreated control. Laser Doppler velo The results of the clinical dermal predictive irritation and sen cimetry (LDV) was used to measure blood flow upon sitization studies on PG described in this section are summar removaL After 3 and 24 hours, the irritation index for PG ized in Table 6. was 1.1 (6 participants) and 1.2 (10 participants), respec tively, indicating a 1-fold increase in blood flow to the test Propylene glycol. A 24-hour single insult occlusive patch test site. The irritation index for oleic acid/PG was 2.1 (6 parti (SIOPT) was performed on an undiluted deodorant formula cipants) and 3.9 (10 participants) after 3 and 24 hours, tion containing 69.15% PG using 20 participants (gender not respectively. Visually, the 24-hour application of PG pro specified).40 A clear stick deodorant was used as a reference duced only slight erythema, while the 24-hour application of control. The test sites were scored on a scale of 0 to 4. With oleic acid/PG produced clearly visible irritation. Distributed for comment only -- do not cite or quote

Fiume et al 2555

Thirty-day use studies were completed with 26 male, 40 reaction. While the authors stated that a stick deodorant for female, and 24 male participants to evaluate the potential for mulation containing 73% PG "did not indicate a clinically deodorant sticks containing 35%,43 65.2%,44 and 73% PG,45 significant potential for dermal irritation or allergic contact respectively, to induce dermal irritation and/or sensitization. sensitization," the Expert Panel questioned that conclusion The participants were instructed to apply the product to the since repeated reactions were observed. underarm once daily for 30 days. None of the participants had Another RIPT was completed with 99 participants to deter any irritation or sensitization reactions, and the researchers mine the sensitization potential of a stick antiperspirant formu concluded that the deodorant sticks containing 35%, 65.2%, lation containing 86% PG.49 (Initially, 113 participants were or 73% PG did not demonstrate a potential for eliciting dermal enrolled in the study; withdrawal was not due to adverse irritation or sensitization. In a 4-week use study completed with effects.) Occlusive patches containing 0.2 g of the test formu 26 male participants following the same procedure, a deodorant lation were applied to the infrascapular region of the back 9 stick containing 65.8% PG also did not demonstrate a potential times during induction and once during challenge. One "+" for eliciting dermal irritation or sensitization.46 reaction was observed during the entire study. There was no evidence of sensitization with an antiperspirant containing Predictive Testing-Sensitization 86%PG. Propylene glycol. A maximization test was completed with 25 participants, 18 male and 7 female, to determine the sensitiza Provocative Testing-Sensitization tion potential of a deodorant containing 69.15% PG.47 During Propylene glycol. Thirty-six patients with chronic venous the induction phase, an occlusive patch containing 0.1 mL of insufficiency (CVI) were patch tested with 5% PG in petro 0.25% aq sodium Iaury! sulfate (SLS) was applied for 24 hours latum by application to the back for 2 days. 5° Twelve patients to the outer arm, volar forearm, or the back of each participant. were male; 2, 5, and 5, had first-, second-, and third-degree That patch was removed and an occlusive patch containing 0.1 CVI, respectively. Twenty-four patients were female; 5 and mL of the test substance was applied to the same site for 48 to 19 had second- and third-degree CVI, respectively. (Proce 72 hours, after which time the patch was removed and the site dural details not provided.) The results were read after 2 and examined. If there was no irritation, the sequence was repeated 3 days; doubtful reactions were read after 4 days. The sensi with the SLS and test article patches for a total of 5 induction tization rate as a percentage of all patients was 8.3%. The exposures. If irritation occurred at any time, the SLS patch was sensitization rate of patients with second- and third-degree excluded. After a 10-day nontreatment period, a challenge was CVI tested with PG was 10% and 8.3%, respectively. Signifi performed in which a previously unexposed site opposite the cant differences were found between males and females; test site was first pretreated with an occlusive patch containing 12.5% of females were sensitized while O% of males were 0.1 mL of 5% aq SLS for 1 hour. Then an occlusive patch sensitized. containing the test substance was applied for 48 hours, and the During the period 2000 to 2004, 308 patients, Ill males and site was scored 1 hour and 24 hours after removal. All the 197 females, with contact dermatitis were patch-tested using scores were 0 for all participants following challenge. No sen the European standard series and some additional chemicals, sitization reactions were seen to a deodorant containing 69.15% including PG.51 Patches were applied to the upper back using PG. Finn chambers that were held in place with Scanpor tape. The An RIPT was completed with 101 participants, 30 male and patches were removed after 2 days, and the sites were evaluated 71 female, to determine the sensitization potential of a stick after 30 minutes and 4 days. Propylene glycol, 5% in petrola deodorant formulation containing 73% PG.48 During the induc tum, did not cause any positive reactions. tion phase, semiocclusive patches containing 0.2 g of the test material were applied to the upper back of each participant for 24 hours, 3 times per week, for a total of 9 applications. The Photoallergenicity first patch was scored (scale of0-4) immediately after removal, Propylene glycol. Over a 2-year period, 30 males and 52 while all others were scored prior to application of the next females with photoallergic contact dermatitis were photopatch patch 24 to 48 hours later. During the induction phase, a score tested with a standard series of sunscreens as well as some of 2 (moderate reaction) resulted in moving the patch to an additional chemicals, including PG. 52 (Dose not given.) The adjacent site, while a second score of 2 or scores of 3 to 4 allergens were applied in duplicate on the back and covered (marked severe) resulted in discontinuation of dosing. The with opaque tape. After 24 hours, the tape was removed, the challenge was performed approximately 2 weeks after the final test sites evaluated, and one set of test sites was irradiated with induction patch using the same procedure but at an adjacent an UV A dose of 5 J/cm2 (using a Daavlin UV A cabinet), giving 2 previously untested site. Challenge sites were scored 24 and 72 an irradiance of 10.4 mW/cm ; this provided a 320 to 400 nm hours after application. Scores of + (barely perceptible or spectrum. The test sites, which were not covered after irradia spotty erythema) to 2, with some dryness, were observed tion, were evaluated 24 and 72 hours later. While some positive throughout the study. Four participants discontinued dosing reactions were observed to other test agents, PG did not pro during the induction phase because of a second moderate duce a photoallergenic or contact allergy response. Distributed for comment only -- do not cite or quote

256S International journal of Toxicology 31 (Supplement 2)

Enhancement of Irritation Effects wt/wt). Two companies reported that the concentration of pro pylene oxide in PPGs used to make finished products is :510 Propylene glycol. The effect of the addition of PG to an iso ppm propylene oxide. propanol vehicle on the irritant reaction of benzoic acid was In 1984, PG was reported to the FDA as being used in 5676 determined in a nonocclusive test using 15 participants, 7 cosmetic formulations at concentrations ofO% to >50%. As of males and 8 females. 53 Benzoic acid in isopropanol was tested 2009, the use of PG has increased significantly, and PG was at concentrations of 31, 62, 125, and 250 mmol/L without PG reported to FDA as being used in 9747 cosmetic formulations. as well as with the addition of 1%, 2%, 5%, 10%, and 25% PG. Concentration of use has also increased, with bath oil/tablet/ The vehicles were also tested. Visual appearance, laser Dop salt preparations containing up to 99% PG and leave-on for pler flowmetry, and skin color (using a Minolta chromameter) mulations, including deodorants, containing up to 73% PG. The were measured at 20, 40, and 60 minutes after application. PPGs are not as widely used as PG, and the maximum reported Propylene glycol enhanced the strength ofthe reactions to 125 concentration is 22%. Tripropylene glycol is used in 8 formu and 250 mmoi/L benzoic acid but not to 31 or 62 mmol/L lations, 7 of which are deodorants, at up to 22%. benzoic acid. (This was observed using all 3 measurement Dermal penetration ofPG from a ternary cosolvent solution methods.) Enhancement was observed with the addition of through hairless mouse skin was 57% over a 24-hour period. 1% PG, and maximal enhancement was attained with 5%. Using TED-FTIR spectroscopy, it appeared that PG molecules No reaction to application of the vehicles was observed. did not reach the dermis in human skin. Propylene glycol can act as a penetration enhancer for some Retrospective Analyses chemicals and under some conditions. Often, it works syner The North America Contact Dermatitis Group (NACDG) per gistically with other enhancers. The mechanism by which PG formed a number of retrospective analyses on various derma enhances penetration has not been definitively identified. tological conditions, determining that :56.0% of patients tested Few toxic effects were seen in dosing with PG or PPGs. In had positive reactions to 30% aq PG. These studies are sum an acute oral toxicity study, 3 of 3 rats dosed with 250 mglkg marized in Table 7. and 2 of 3 rat dosed with 1000 mglkg PPG 425 survived. All mice survived in a repeated-dose oral toxicity study in which mice were given 10% PG in drinking water for 14 days. Case Reports Repeated-dose inhalation data reported some effects in rats due A few case reports have been described concerning PG and to PG exposure of 2.2 mg/L air for 6 hoursld, 5 days/week, for hand dermatitis or atopic dermatitis. Patch test results generally 13 weeks, but these effects were inconsistent and without dose had a positive reaction to PG in these case studies. Improve response trends. ment was seen with the avoidance of PG-containing Undiluted PG was less than marginally irritating to rabbit 54 55 products. • eyes In a dermal irritation study in mice, undiluted PG was minimally irritating. Summary Oral administration of PG did not have any adverse repro ductive or developmental effects when evaluated in mice and Both PG and PPGs were reviewed by the CIR Expert Panel in rats at doses of :51600 mglkg, rabbits at doses of :51230 mglkg, 1994, and it was concluded that these ingredients were safe for or hamsters at doses of :51550 mglkg. A study examining use in cosmetic products at concentrations up to 50.0%. This induction of cytogenetic aberrations in mice reported an rereview was opened to amend the conclusion (the concentra increase in the frequency of premature centrosphere separation tion of use ofPG is >50%), consider new data, and to add new with 1300 to 5200 mglkg PG. In zygotes from PG-dosed mice, ingredients so that all of the PPGs identified in the Interna hyperploidy was increased. tional Cosmetic Ingredient Dictionary and Handbook, that is, Tripropylene glycol, :510 000 Jlg/plate, was not mutagenic PPG-3, PPG-7, PPG-9, PPG-12, PPG-13, PPG-15, PPG-16, in an Ames assay. PPG-17, PPG-20, PPG-26, PPG-30, PPG-33, PPG-34, Combined exposure to PG and oleic acid synergistically PPG-51, PPG-52, and PPG-69, as well as tripropylene glycol, enhanced the dermal penetration of both compounds. Addi are included. tion of PG to an isopropanol vehicle enhanced the irritant Propylene glycol is an aliphatic alcohol that is manufactured reactions of benzoic acid; maximal enhancement was seen as a reaction product of propylene oxide and water. Tripropy with 5% PG. lene glycol is manufactured by sequential addition of propylene The dermal irritation potentials of deodorant formulations oxide to PG and is only of 3 chain length. Polypropylene gly containing 68.06% or 69.15% PG were evaluated in an SIOPT cols are manufactured by the addition of propylene oxide to and compared to a reference in-use control formulation; the dipropylene glycol and have average chain lengths of their "n" formulations containing PG were no more irritating or even value; for example, PPG-3 would have an average chain length less irritating than the reference control. Use studies of of 3. The USP grade PG (used in cosmetics) manufactured by deodorant formulations containing 35% to 73% PG did not Dow contains diethylene glycol and ethylene glycol at concen report any potential for eliciting irritation or sensitization. trations that are nondetectable (quantification limit of 0.008% Deodorant formulations containing 69.15% or 86% PG did Distributed for comment only -- do not cite or quote

Fiume et al 257S

Table 7. Retrospective Analyses With Propylene Glycol

Years No. of Patients Studied %PG Methods Findings

Not given 1984-1996 10 aq. data were collected from NACDG- the SPIN rank for PG has changed over reported studies; the SPIN for each time: 23 in 1984-1985; 40 in 1992- allergen was calculated as the pro- 1994; 41 in 1994-199666 portion of the population allergic by the weighted clinician-assessed likelihood of relevance of the reaction 45 138 patients 1992-2002 20 aq. Analysis of a large pool of IVDK patch- - I044 patients (2.3%), 412 males and (16 210 males; test data, examining possible rele- 632 females, had positive reactions; 895, 28 928 females) vance of patient characteristics 129, and 20 patients had I+. 2+. and 3+ reactions, respectively; of the 895 I + reactions, I 14 were to PG only; - 1041 doubtful, 43 follicular, and 271 irritant reactions were observed; - there were little difference between patients with positive and negative reactions to PG; the greatest difference was the high portion (2 7.2% vs. 13.1 %) of patients with leg dermatitis - this was the only sig. risk factor; - the most common concomitant reactions were with fragrance mix, balsam of Peru, lanolin alcohol, amerchol l-1 0 I. and nickel sulfate67 23 359 patients 1996-2006 30 aq. retrospective cross-sectional analysis of - 810 patients (3.$%) had reactions to NACDG patch-test data to evaluate PG; 12.8% of the reactions were the patient characteristics, clinical definitely relevant, 88.3% were relevance (definite- positive reaction currently relative (definite, probable or to a PG-containing item; probable- possible relevance), 4.2% were PG was present in the skin contac- occupation related; tants; possible-skin contact with -135 patients were positive to only PG; PG-containing material was likely), in these patients, the face was the most source of exposure, and occupational commonly-affected area (25.9%), a relationship scattered or generalized pattern was next (23.7%); - the most common concomitant reactions were with balsam of Peru, fragrance mix, formalde~de, nickel sulfate, and bacitracin 1494 patients 2001-2004 30 aq. retrospective analysis of cross-sectional 89 patients (6.0%) had positive w/SGD (patient NACDG data using only patients reactions to PG; 94% of the reactions pop. 10 061) with SGD as the sole site affected were currently relative, with 30.3, 20.2, and 42.7% being of definite, probable, and possible relevance'' I0 061 patients 2001-2004 30 aq. retrospective analysis of cross-sectional I 09 patients ( 1.1%), 37 males and n NACDG data to determine reactions females, had 122 reactions to foods; of to foods those 122 reactions, 5 were to PG 70

Abbreviations: IVDK, Information Network of Departments of Dermatolou; NACDG, North America Contact Dermatitis Group; SGD, scattered generalized distribution; SPIN, significance-prevalence index number

not induce sensitization reactions; however, questionable a patch test. Retrospective analysis of pools of patient patch results were obtained in an RIPT of a deodorant containing test data, mostly NACDG data, indicated that, 6.0% or less 73% PG. of patients tested had positive reactions to 30% aq PG. In a provocative study in which the sensitization poten Additionally, PG (concentration not speci fled) did not pro tial of PG was evaluated in patients with contact dermatitis, duce a photoallergic response in a provocative photopatch 5% PG in petrolatum did not cause any positive reactions in test. Distributed for comment only -- do not cite or quote

258S International journal of Toxicology 31 (Supplement 2)

Discussion "n" is 120, the ingredient would be sufficiently large so that no dermal penetration would be possible. The CIR Expert Panel reopened the 1994 safety assessment of PG and PPGs to address the safety of current high-use concentrations of PG as well as to add all the PPGs currently Conclusion listed in the International Cosmetic Ingredient Dictionary and The CIR Expert Panel concluded that PG, tripropylene glycol, Handbook. This report is intended to also address the safety of PPG-3, •7, -9, -12, -13, -15, -16, -17, 20, -26, -30, -33, -34, -51, similar PPGs that may be used as cosmetic ingredients in the -52, -69, and any PPG ;:::3 are safe as cosmetic ingredients in future. the present practices of use and concentration as described in Since tripropylene glycol is similar to PG and the PPGs, its this safety assessment when formulated to be nonirritating. safety can be supported by the existing data and, therefore, the Were ingredients in this group not in current use to be used Panel included tripropylene glycol in this safety assessment. in the future, the expectation is that they would be used in Propylene oxide is used in the manufacture of PPGs but product categories and at concentrations comparable to others should not appear in cosmetic formulations because of safety in the group. concerns. The Panel expects that PPGs contain ::; 10 ppm pro pylene oxide, ensuring the safety of formulations in which Authors' Note PPGs are used. Unpublished sources cited in this report are available from the Direc Both PG and PPGs were not considered to be acute or tor, Cosmetic Ingredient Review, 1101 17th St, Suite 412, Washing chronic toxicants in oral or dermal studies, were not genotoxic ton, DC 20036, USA. or carcinogenic, and were not reproductive or developmental toxicants, supporting that their use in cosmetics would be safe Declaration of Conflicting Interest in regard to these end points. The author(s) declared no potential conflicts of interest with respect to At the time of the original safety assessment, a concentra the research, authorship, and/or publication of this article. tion limit of 50% PG and PPGs was established based on the results of existing irritation and sensitization studies. The Funding potential for skin irritation was especially of concern under The author(s) disclosed receipt of the following financial support for occlusive conditions, and this potential could be concentration the research, authorship, and/or publication of this article: The articles dependent. An RIPT performed using a stick antiperspirant in this supplement were sponsored by the Cosmetic Ingredient containing 86% PG produced no evidence of sensitization. Review. The Cosmetic Ingredient Review is financially supported Additionally, use studies of deodorant sticks containing 35% by the Personal Care Products Council. to 73% PG did not demonstrate a potential for eliciting dermal irritation or sensitization. Therefore, the Panel determined that References PG would not present a sensitization risk at the concentrations 1. Elder RE. Final report on the safety assessmento f butylene gly currently in use. col,. hexylene glycol, ethoxydiglycol, and dipropylenen glycol. The Expert Panel did note that PG may act as a penetration JAm Coli Toxicol. 1985;4(5):223-248. enhancer. Some cosmetic ingredients have been regarded as 2. Andersen FA. Annual review of cosmetic ingredient safety safe based on the fact that they do not penetrate the skin. If assessments - 2004/2005. Int J Toxicol. 2006;25(suppl 2): 1-89. PG enhances penetration of such ingredients, then they should 3. Gottschalck TE, Bailey JE. International Cosmetic Ingredient not exist together in formulation. Dictionary and Handbook. 13th ed. Washington, DC: Personal Additionally, PG is used in aerosols. The potential adverse Care Products Council; 2010. effects of inhaled aerosols depend on the specific chemical 4. Hazardous Substances Data Bank (HSDB). Tripropylene glycol • species, the concentration and the duration of the exposure, Method of manufacture. http://toxnet.nlm.nih.gov/cgi-bin/sis/ and their site of deposition within the respiratory system. In search/f'Utemp/ -pfrkn6: I. Accessed October 7, 2009. practice, aerosols should have at least 99% of their particle 5. Andersen FA. Final report on the safety assessment of propylene diameters in the 10 to II 0 J.lm range and the mean particle glycol and polypropylene glycols. JAm Coli Toxicol. 1994; 13(6): diameter in a typical aerosol spray has been reported as 437-491. "'3 8 Jlm. Particles with an aerodynamic diameter of :::;; 10 J.lm 6. Dow Chemical Company. Letter regarding USP standards dated are respirable. In the absence of significant inhalation toxicity Jan 29, 2010. http:/fwww.dow.com/propyleneglycollletter.pdf. data, the Panel determined that PG can be used safely in hair Accessed February 21,2010. sprays because the product particle size is not respirable. 7. Personal Care Products Council (The Council). Concentration of The CIR Expert Panel, as noted earlier, considers that the propylene oxide in polypropylene glycol ingredients. Unpub available data for PPG-3 through PPG-69 would extend to any lished data received from the Council. Apri12010. PPG-n to be used in cosmetics in the future. There are no 8. Food and Drug Administration (FDA). Frequency of use of cos concerns regarding residual monomers in PPGs. If the "n" in metic ingredients. FDA Database. Washington, DC: FDA; 2009. PPG-n is 32, for example, ample evidence suggests that its 9. Personal Care Products Council (The Council). Concentration of toxicity would be no different from PPG-30 or PPG-33. If the Use- Propylene Glycol, PPG-3, PPG-7, PPG-9, PPG-12, PPG-13, Distributed for comment only -- do not cite or quote

Fiume et al 2595

PPG-15, PPG-16, PPG-17, PPG-20, PPG-26, PPG-30, PPG-34, 28. Morshed KM, Desjeux JF, Nagpaul JP, Majumdar S, Amma MK. PPG-51, PPG-52, PPG-69, Tripropylene Glycol. Originally pre The effect of propane-dials on the intestinal uptake of nutrients pared Sept 15; updated Dec 3. Unpublished data received from the and brush border membrane enzymes in the rat. Biocltem Med Council. September 2009. Metab Bioi. 1991;45(2):161-170. I 0. James AC, Stahlhofen W, Rudolf G, et a!. Deposition of inhaled 29. Liesivuori J, Laitinen J, Savolainen H. Rat model for renal effects particles. Ann ICRP. 1994;24(1-3):231-232. of 2-alkoxyalcohols and their acetates. Arch Toxico/. 1999; II. Oberdorster G, Oberdorster E, Oberdorster J. Nanotoxicology: an 73(4-5):229-232. emerging discipline evolving from studies of ultrafine particles. 30. Environmental Health Research and Testing. Propylene glycol: Environ Health Perspect. 2005;113(7):823-839. reproduction and fertility assessment in CD-I mice when admi 12. Bower D. 1999. CIR Expert Panel meeting. Unpublished infor nistered in drinking water. (Revised September 1985.). NTIS No. mation on hair spray particle sizes provided at the September PB86140662. 1985. 1999. 31. Suber RL, Deskin R, Nikiforov I, Fouillet X, Coggins CRE. 13. Jolmson MA. The influence of particle size. Spray Technology Subchronic nose-only inhalation study of propylene glycol and Marketing. 2004:24-27. in Sprague-Dawley rats. Food Chem Toxicol. 1989;27(9): 14. European Commission. Coslng. Cosmetic Ingredients and Sub 573-583. stances. http://ec. europa.eulenterpriselcosmetics/cosing/index. 32. Shirwaikar A, Gundu, Rao P. Ophthalmic irritation potential of cfm?fuseaction=search.resu/ts. Accessed 9, 2009. propylene glycol. Indian J Pharm Sd. 1995;57(3):109-112. 15. Ministry of Health, Labour, and Welfare (MHLW). Standards for 33. Phillips CA, Michniak BB. Topical application of Azone analogs cosmetics. MHLW Notification No. 331. http://www.mhlw.gojp/ to hairless mouse skin: Histopathological study. lnt J Pharm. englishltopics/cosmetics/index.html. Accessed October 19, 2009. 1995;125:63-71. 16. Code of Federal Regulations. Title 21. Food and Drugs. Revised 34. Food and Drug Research Labs, Inc. Teratologic evaluation of April I, 2009. http://www.access.gpo.gov/cgi-bin/cfrassemble. FDA 71-56 (propylene glycol). Final report dated July 31. NTIS cgi. Accessed January II, 20 I 0. Publication No.PB223822. July 31st, 1973. 17. FDA. Propylene Glycol. Monograph prepared by the FDA Center 35. Minck DR, Acuff-Smith KD, Vorhees CV. Comparison of the for Food Safety and Applied Nutrition. behavioral teratogenic potential of phenytoin, mephenytoin, etho 18. Ministry of Health, Labour, and Welfare (MHLW). Notification toin, and hydantoin in rats. Teratology. 1991 ;43(4):279-293. No. 498. http ://www.mhlw. go.jp/english/topics/foodsafety/ 36. National Toxicology Program. Salmonella study summary. Study positivelist060228/dl/no2.pdf. Accessed October 19, 2009. A04474. Tripropylene glycol. http://ntp-apps.niehs.nih.gov/ntp_ 19. Food and Drug Administation. Inactive Ingredient Search for tox/index.cfm?fuseaction=salmonella.salmonellaData&study_ Approved Drug Products. http:l/www.accessdata.fda.gov/scripts/ no=A044 74&cas_no- 24800%2D44 %2DO&endpointlist=SA. cder/iiglgetiigWEB.cfm. Accessed April4, 2010. Accessed October 7, 2009. 20. National Toxicology Program. CAS Registry Number: 57-55-6. 37. Schneider IM, Dohner B, Neubert R, Wohlrab W. Evaluation of Information from ChemiDPius and HSDB. http://ntp.niehs.nih. drug penetration into human skin ex vivo using branched fatty gov/index.cfm?objectid=E880F4CO-BDB5-82F8-FBE707DE acids and propylene glycol./nt J Pharm. 1996;145:187-196. D5FF5682. Accessed September 7, 2009. 38. Tanojo H, Junginger HE, Bodde HE. In vivo human skin perme 21. Lewis RJ Sr. Hawley's Condensed Chemical Dictionary. 13th ed. ability enhancement by oleic acid: transepidermal water loss and New York, NY: John Wiley & Sons, Inc.; 1997. Fourier-transform infrared spectroscopy studies. J Controlled 22. National Toxicology Program. CAS Registry Number: 24800-44- Release. 1997;47:31-39. 0 (tripropylene glycol). http://ntp.niehs.nih.gov/index.cfm? 39. Funk JO, Maibach HI. Propylene glycol dermatitis: re-evaluation objectid=E8 7 EDC5 9-BDB 5-82F8- FC884A 68 60B5FOC6. of an old problem. Contact Dermatitis. 1994;31(4):236-241 . Accessed October 7, 2009. 40. Menning M. Human single insult occlusive patch test with a stick 23. Squillante E, Needham T, Maniar A, Kislalioglu S, Zia H. Codif deodorant containing 69.15% propylene glycol. Study dated fusion of propylene glycol and dimethyl isosorbide in hairless 1997. Unpublished data received from the Council. Submitted mouse skin. Eur J Pharm Biopharm. 1998;46(3):265-271. September 2009. 24. Notingher I, Imhof RE. Mid-infrared in vivo depth-profiling of 41. Menning M. Human single insult occlusive patch test of a deo topical chemicals on skin. Skin Res Techno/. 2004;10(2):113-121. dorant containing 68.06% propylene glycol. Study dated 1998. 25. Ponec M, Haverkort M, Soei YL, Kempenaar J, Bodde H. Use of Unpublished data received from the Council. Submitted Septem human keratinocyte and fibroblast cultures for toxicity studies ber 2009. of topically applied compounds. J Pharm Sci. 1990;79(4): 42. Tanojo H, Boelsma E, Junginger HE, Ponce M, Bodde HE. In 312-316. vivo human skin permeability enhancement by oleic acid: laser 26. Dow Chemical Co. Initial submission: letter regarding an oral Doppler velocimetty study. J Controlled Release. 1999;58: gavage study in rats dated 041993. 1993. Report No. DOCNO 97-104. TSCATS /424506. 43 . Clinical Research Laboratories, Inc. Deodorant use study on a 27. Mailhes JB, Young D, London SN. 1,2-Propanediol-induced pre deodorant stick containing 35% propylene glycol. CRL study mature centromere separation in mouse oocytes and aneuploidy in no. CRL35106. Unpublished data received from the Council. one-cell zygotes. Bioi Reprod. 1997;57(1):92-98. Submitted January 2010. Distributed for comment only -- do not cite or quote

2605 International journal of Toxicology 31 (Supplement 2)

44. Clinical Research Laboratories, Inc. Deodorant use study on a %2FChemFull.jsp%3FcalledFrom%3Dlite&chemid=024800440 deodorant stick containing 65.2% propylene glycol. CRL Study &formatType=_3D. Accessed October 7, 2009. No. CRL 1231 OS. Study dated Jan II, 2006. Unpublished data 58. Food Chemicals Codex. 4th ed. Washington, DC: National Acad received from the Council. Submitted January 2010. emy Press; 1996. 45. Clinical Research Laboratories, Inc. Final report: deodorant use 59. Food Chemicals Codex. 3rd ed. National Academy of Sciences; study on a deodorant stick containing 73% propylene glycol. CRL 1981. study no. CRL112804. 2004. Unpublished data received from the 60. Sax Nl. Dangerous Properties ofIndustrial Materials. New York, Council. Submitted September 2009. NY: Van Nostrand Reinhold; 1979. 46. Clinical Research Laboratories, Inc. Four week safety in-use 61. Niazy EM, Molokhia AM, El-Gorashi AS. Effect of vehicle and study with a deodorant stick containing 65.8% propylene glycol. drug concentration on transdermal delivery of dihydroergotamine CRL study no. CRLI07206. Study dated Nov 2, 2006. Unpub using excised animal skin. Drug Dev lnd Ph arm. 1990; 16: lished data received from the Council. Submitted January 2010. Pl697-P1715. 47. KGL, Inc. The dermination of the contact-sensitizing potential of 62. Bendas B, Schmalfuss U, Neubert R. Influence of propylene gly one material (a deodorant containing 69.15% propylene glycol) by col as coso!vent on mechanisms of drug transport from hydrogels. means of the maximization assay. Study dated 1997. Unpublished lnt J Pharm. 1995;116:19-30. data received from the Council. Submitted September 2009. 63. Levang AK, Zhao K, Singh J. Effect of ethanol/propylene glycol 48. Consumer Product Testing Co. Final report: repeated insult patch on the in vitro percutaneous absorption of aspirin, biophysical test of a deodorant stick containing 73% propylene glycol. Experi changes and macroscopic barrier properties of the skin. lnt J ment ref. no. C04-1274.02. Study dated 2005. Unpublished data Pharm. 1999;( 181 ):255-263. received from the Council. Submitted September 2009. 64. Arellano A, Santoyo S, Martin C, Ygartua P. Influence of propy 49. TKL Research. Human repeated insult patch test with an antiper lene glycol and isopropyl myristate on the in vitro percutaneous spirant containing 86% propylene glycol. TKL Study No. penetraton of diclofenac sodium from carbopol gels. Eur J Pharm DS104808-t. Study reissued Jan 27, 2010. Unpublished data Sci. 1999;7(2):129-135. received from the Counci I. Submitted February 20 I 0. 65. Clinical Research Laboratories, Inc. Final report: Deodorant use 50. Gallenkemper G, Rabe E, Bauer R. Contact sensitization in study on a deodorant stick containing 73% propylene glycol. CRL chronic venous insufficiency: modem wound dressings. Contact study no. CRLI12804. 2004. Unpublished data received from the Dermatitis. 1998;38(5):274-278. Council. Submitted September 2009. 51 . Boyvat A, Akyol A, rgey E. Contact sensitivity to preservatives in 66. Maouad M, Fleischer AB, Jr, Sherertz EF, Feldman SR. Turkey. Contact Dermatitis. 2005;52(6):329-332. Significance-prevalence index number: a reinterpretation and 52. Rodriguez E, Valbuena MC, Rey M, Porras de Quintana L. Causal enhancement of data from the North American contact dermatitis agents of photoallergic contact dermatitis diagnosed in the group. JAm Acad Dermatol. 1999;41(4):573-576. national institute of dermatology of Colombia. Photodermatol 67. Lessmann H, Schnuch A, Geier J, Uter W. Skin-sensitizing and Photoimmunol Photomed. 2006;22(4): 189-192. irritant properties of propylene glycol. Contact Dermatitis. 2005; 53. Lahti A, Noponen SL. Propylene glycol in an isopropanol vehicle 53(5):247-259. enhances immediate irritant reactions to benzoic acid. Contact 68. Warshaw EM, Botto NC, Maibach HI, et al. Positive patch-test Dermatitis. I 998;39(3): 150-151. reactions to propylene glycol: a retrospective cross-sectional anal 54. Lu R, Katta R. Iatrogenic contact dermatitis due to propylene ysis from the North American contact dermatitis group, 1996 to glycol. J Drugs Dermatol. 2005;4(1):98-101. 2006. Dermatitis. 2009;20(1): 14-20. 55 . Lowther A, McCormick T, Nedorost S. Systemic contact derma 69. Zug KA, Rietschel RL, Warshaw EM, et al. The value of patch titis from propylene glycol. Dermatitis. 2008;19(2):105-108. testing patients with a scattered generalized distribution of der 56. National Institute for Occupational Safety and Health (NIOSH). matitis: retrospective cross-sectional analyses of North American International Chemical Safety Card- Tripropylene Glycol. http:// Contact Dermatitis Group data, 2001 to 2004. JAm Acad Derma www.cdc.gov/niosh/ipcsneng 1348.html. Accessed October 7, to/. 2008;59(3):426-431. 2009. 70. Warshaw EM, Botto NC, Zug KA, et al. Contact dermatitis asso 57. ChemiDPlusLite. Tripropylene glycol. http://chem.sis.nlm.nih. ciated with food: retrospective cross-sectional analysis of North gov/chemidplus/ProxyServlet?objectHandle=Search&action American contact dermatitis group data, 2001-2004. Dermatitis. Handle=getAII3DMViewFiles&nextPage=jsp%2Fcommon 2008; 19(5):252-260. Distributed for comment only -- do not cite or quote

International journal of ToxicolocY 31(Supplement 2) 2615 ~ The Aulhor(s) 2012 Erratum Reprints and permission: ~a~epub.comljoumalsPermissions.nav DOt: IO. IIn/1091581812464367 hctp:/f~t.sagepub .com @SAGE

Fiume MM, Heldreth B, Bergfeld WF, Belsito DV, Hill RA, Klaassen CD, Liebler D, Marks JG, Shank RC, Slaga TJ, Snyder PW, Anderson FA. Final Report ofthe Cosmetic Ingredient Review Expert Panel On the Safety Assessment ofDicarboxylic Acids, Salts, and Esters./ntJ Toxicol. 2012;31(4S): 5S-76S. (Original doi: 10.1177/ 1091581812447203)

The second author should have been listed as Bart Heldreth. Distributed for comment only -- do not cite or quote

ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 07F - Makeup Bases 1 ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 08A - Basecoats and Undercoats 35 ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 08C - Nail Creams and Lotions 4 ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 08D - Nail Extenders 1 ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 08E - Nail Polish and Enamel 348 ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 08F - Nail Polish and Enamel Removers 1 ADIPIC ACID/NEOPENTYL GLYCOL/TRIMELLITIC ANHYDRIDE COPOLYMER 08G - Other Manicuring Preparations 21

PHTHALIC ANHYDRIDE/TRIMELLITIC ANHYDRIDE/GLYCOLS COPOLYMER 07I - Other Makeup Preparations 2 PHTHALIC ANHYDRIDE/TRIMELLITIC ANHYDRIDE/GLYCOLS COPOLYMER 08A - Basecoats and Undercoats 13 PHTHALIC ANHYDRIDE/TRIMELLITIC ANHYDRIDE/GLYCOLS COPOLYMER 08C - Nail Creams and Lotions 1 PHTHALIC ANHYDRIDE/TRIMELLITIC ANHYDRIDE/GLYCOLS COPOLYMER 08D - Nail Extenders 1 PHTHALIC ANHYDRIDE/TRIMELLITIC ANHYDRIDE/GLYCOLS COPOLYMER 08E - Nail Polish and Enamel 50 PHTHALIC ANHYDRIDE/TRIMELLITIC ANHYDRIDE/GLYCOLS COPOLYMER 08G - Other Manicuring Preparations 7 Distributed for comment only -- do not cite or quote

Concentration of Use by FDA Product Category – Trimellitic Anhydride Copolymers*

Adipic Acid/Neopentyl Glycol/Trimellitic Anhydride Copolymer Adipic Acid/CHDM/MA/Neopentyl Glycol/Trimellitic Anhydride Copolymer Isostearoyl Trimellitic Anhydride/Trimethylolpropane Copolymer Phthalic Anhydride/Trimellitic Anhydride/Glycols Copolymer Propylene Glycol/Sebacic Acid/Trimellitic Anhydride Copolymer TDI/Trimellitic Anhydride Copolymer Trimethylpentanediol/Isophthalic Acid/Trimellitic Anhydride Copolymer

Ingredient Product Category Maximum Concentration of Use Adipic Acid/Neopentyl Glycol/Trimellitic Basecoats and undercoats 9-9.7% Anhydride Copolymer (manicuring preparations) Adipic Acid/Neopentyl Glycol/Trimellitic Nail creams and lotions 6.2% Anhydride Copolymer Adipic Acid/Neopentyl Glycol/Trimellitic Nail extenders 22% Anhydride Copolymer Adipic Acid/Neopentyl Glycol/Trimellitic Nail polish and enamel 5.4-32.8% Anhydride Copolymer Adipic Acid/Neopentyl Glycol/Trimellitic Other manicuring preparations 18.2% Anhydride Copolymer Adipic Acid/Neopentyl Glycol/Trimellitic Face and neck products Anhydride Copolymer Not spray 1%

Phthalic Anhydride/Trimellitic Basecoats and undercoats 6.9% Anhydride/Glycols Copolymer (manicuring preparations) Phthalic Anhydride/Trimellitic Nail polish and enamel 1.8-12% Anhydride/Glycols Copolymer Phthalic Anhydride/Trimellitic Other manicuring preparations 6% Anhydride/Glycols Copolymer

TDI/Trimellitic Anhydride Copolymer Mascara 1%

*Ingredients found in the title of the table but not included in the table were included in the concentration of use survey, but no uses were reported.

Information collected in 2015 Table prepared April 9, 2015 Distributed for comment only -- do not cite or quote

Distributed for comment only -- do not cite or quote